{"status":"ok","feed":{"url":"https://newsletter.kiin.bio/feed","title":"Kiin Bio Weekly","link":"https://newsletter.kiin.bio/","author":"Kiin Bio","description":"Where AI meets Life Science","image":"https://substackcdn.com/image/fetch/$s_!UiEF!,w_256,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8654bb64-0b90-4220-9c12-7c9269dd2c95_1093x1093.png"},"items":[{"title":"\ud83e\uddecSable Bio: Building the Safety Layer Drug Discovery Has Been Missing","pubDate":"2026-05-05 17:01:57","link":"https://newsletter.kiin.bio/p/sable-bio-building-the-safety-layer","guid":"https://newsletter.kiin.bio/p/sable-bio-building-the-safety-layer","author":"Natasha Kilroy","thumbnail":"","description":"Deep Dive | Edition 18","content":"\n
Welcome back to the deep dive, where we break down the AI tools and data reshaping how new drugs are discovered. In each edition, we speak directly with the teams behind these tools to explain what they solve, how they work and where they are going next.
\n
\n
Keeping up with AI x life science news can get exhausting.
\n
It\u2019s scattered across LinkedIn, X, Substack, arXiv, Slack, newsletters... and you still somehow miss the things that actually matter. Too much noise, not enough signal.
\n
We\u2019re building something to fix that: a smarter, more powerful way to stay on top of what\u2019s actually relevant to you.
\n
But we want to build it with you, not just for you. Take 2 minutes to tell us what\u2019s missing. What you share will directly shape what we build, and you\u2019ll be the first to benefit from it.
Around 30% of clinical trials fail due to safety concerns. That\u2019s billions in sunk costs, years of lost time, and patients who don\u2019t get the medicines they need. Sable Bio thinks the problem starts much earlier in the pipeline, with how safety assessment is done in the first place.
I spoke with Olly Oechsle, CTO of Sable Bio, about how time-consuming traditional safety workflows can be and how the company\u2019s Target Intelligence platform helps toxicologists.
\n
\n
\ud83d\udd34 The Problem
\n
Before a drug candidate moves into preclinical development, safety scientists need to answer a deceptively simple question: if we inhibit or activate this target, what else is going to happen that we didn\u2019t intend?
\n
Answering that means looking across clinical trial data, genetic association studies, mouse knockout phenotypes, scientific literature, expression profiles, and more. Each lives in a different database, requires different expertise to interpret, and few are built with a toxicologist\u2019s specific needs in mind.
\n
The result is a process that takes anywhere from a few days to a month. Safety scientists spend a lot of that time collating information rather than doing the analytical work they\u2019re trained for: identifying risks, building mitigation strategies, and making judgment calls about whether a target\u2019s safety profile balances its therapeutic potential.
\n
There\u2019s also a reproducibility problem. \u201cIf another scientist were to do the same research with the same amount of time, would they have come up with the same answer?\u201d Olly asks. When you\u2019re manually searching PubMed, reviewing mouse knockouts or sifting through clinical databases, you inevitably go deep on some rabbit holes while missing others entirely. There\u2019s no systematic way to know whether every potential adverse event has been examined across every relevant data source.
\n
And the challenge doesn\u2019t end with a single assessment. Drug discovery programs run for years. New data emerges, new papers are published, and safety scientists, often stretched across multiple projects, need to stay on top of all of it.
\n
\n
\ud83d\udca1 The Platform
\n
Sable\u2019s core product is Target Safety Reports. A user searches for any target in the genome, specifies whether they plan to inhibit or activate it, and receives a comprehensive, customized safety report.
The report overview for a DGAT1 inhibitor, with Sable\u2019s Risk Radar summarising safety signals across data sources at a glance.
\n
Findings can be organized by organ system (cardiovascular risk, hepatic risk, and so on) or by data source, letting users drill into the literature, clinical data, genetic associations, or expression profiles separately. The platform pulls from PubMed, clinical trial databases, GWAS data, gene burden studies, and MGI mouse knockout data, among other sources.
\n
What makes this more than a search engine is how Sable has tuned each data source specifically for safety science and collated the result into a coherent report. For clinical trial data, statistical methods distinguish drug-centric effects from target-driven ones, and on-target from off-target effects, accounting for patient comorbidities that can be mistaken for causative drug effects. For the literature, Sable has built proprietary language models that extract target-to-adverse-event relationships with precision and coverage that general-purpose LLMs can\u2019t match.
\n
\u201cIt\u2019s easy to make early strides with literature but it\u2019s really hard to do it well,\u201d Olly says. General AI models can answer questions about a single paper, but they struggle with corpus-wide analysis, and tend to \u201centhusiastically offer insights\u201d from their broader training data that goes beyond the evidence being reviewed. Meanwhile literature tools can quickly present an overwhelming volume of content.
A target safety report broken down by organ system, bringing together mouse knockout phenotypes, expression data, and biological function signals in a single view.
\n
The platform also includes tracking that alerts users when new information emerges affecting a target\u2019s safety profile, plus collaboration features for commenting, discussing, and sharing reports across teams.
\n
\n
\ud83d\udcca Weight of Evidence
\n
Central to Sable\u2019s approach is the toxicology principle of \u201cweight of evidence.\u201d Rarely does a single data source give a definitive answer on target safety. Instead, signals come from multiple directions: a suggestive mouse knockout phenotype, a mechanism described in the literature, expression data showing the target is active in a particular tissue, and adverse events observed with a related ligand.
\n
Sable brings all of these signals together, letting scientists evaluate the full picture rather than chasing individual threads across separate databases. This works both ways: sometimes the different evidence types show a perceived hazard isn\u2019t actually a concern, potentially saving organizations from expensive and unnecessary preclinical studies.
\n
\n
\ud83e\udde9 Where It Fits
\n
Sable\u2019s sweet spot is late discovery through to lead optimization, where the question shifts from \u201cwill this work?\u201d to \u201cwhat else is this doing?\u201d The platform is used by preclinical safety scientists at several top-10 pharma companies, biotechs, venture capital firms conducting asset due diligence, and by consultants running target safety assessments.
\n
The team is exploring expansion into earlier target selection, and for organizations using AI-driven target ID platforms that generate dozens or hundreds of candidates, Sable is developing wider-scale safety assessment products, systematically evaluating 50 to 200 targets with a traceable decision-making record. These are planned for release within the year.
\n
Beyond the web interface, Sable offers API access and is building MCP integrations, positioning itself as a universal safety layer that plugs into the broader drug discovery ecosystem.
\n
\n
\ud83d\udd2e The Future
\n
Sable is rolling out its new proprietary literature models alongside deeper analysis for single-cell data, expression data, and mechanistic biology. The company is also running a side-by-side comparison study between toxicologist assessments and platform outputs, with results expected in a couple of months.
\u201cWe\u2019re not there to replace the decision-making process, which is ultimately a human one,\u201d Olly says. \u201cWe\u2019re there to save experts\u2019 time.\u201d Given how much of that time currently goes toward collecting data rather than acting on it, that\u2019s a proposition most safety scientists can get behind.
\n
\ud83e\uddd1\u200d\ud83d\udd2cGet in touch with Olly.
We\u2019re always looking to grow our community. If you\u2019d like to get involved, contribute ideas or share something you\u2019re building, fill out this form or reach out to me directly.
Welcome back to your weekly dose of AI news for Life Science!
\n
\n
Keeping up with AI x life science news can get exhausting.
\n
It\u2019s scattered across LinkedIn, X, Substack, arXiv, Slack, newsletters... and you still somehow miss the things that actually matter. Too much noise, not enough signal.
\n
We\u2019re building something to fix that: a smarter, more powerful way to stay on top of what\u2019s actually relevant to you.
\n
But we want to build it with you, not just for you. Take 2 minutes to tell us what\u2019s missing. What you share will directly shape what we build, and you\u2019ll be the first to benefit from it.
\ud83d\udd2c Designing functional enzymes from scratch remains one of the hardest challenges in protein design. Previous approaches relied on borrowing catalytic motifs from nature, but optimising for structure alone does not guarantee catalytic competence. Efficient catalysis requires a precise balance of chemical, geometric, and electrostatic criteria that existing methods struggle to jointly satisfy.
\n
Gina El Nesr and colleagues at Stanford present dEVA (design by EVolutionary Algorithm), a multi-objective framework built on NSGA-II that simultaneously optimises multiple design objectives, enriching for candidates where all criteria are mutually compatible rather than traded off against one another. Using LigandMPNN for sequence design and Metal3D for metal site prediction, dEVA iterates mutations across generations, converging on Pareto-optimal solutions.
\n
\ud83e\uddec Their best design, desB, achieves catalytic efficiency of 1,500 M\u207b\u00b9s\u207b\u00b9 and rate enhancement of 3x10\u00b9\u00b3 without directed evolution. Analysis of PDB zinc sites revealed around 10% had zero coordinating ligands and over 54% had two or fewer residues, many being crystallisation artefacts. They trained Metal3D-Clean and Metal3D-Cat on curated data to address this.
\n
\u26a1 desB hydrolyses both phosphomonoesters (uncatalysed half-life >500,000 years) and phosphodiesters (>13 million years). This promiscuity mirrors early enzyme evolution, opening the door to engineering specificity from a designed starting point.
Design dEVA designed a catalytically active metalloenzyme without natural templates or evolutionary information. No directed evolution was needed to achieve function.
\n
2\ufe0f\u20e3 Multi-Objective Optimisation Over Single-Score Ranking
\n
By treating design as population-based evolutionary search across multiple objectives, dEVA avoids the compromises of single-score optimisation or sequential filtering.
\n
3\ufe0f\u20e3 Training Data Quality Matters
\n
Curating PDB metal sites and retraining Metal3D on catalytically relevant examples was essential. Garbage in, garbage out applies to structural prediction too.
\n
4\ufe0f\u20e3 A Platform for Designed Catalysis
\n
The promiscuous activity of desB mirrors early enzyme evolution, providing a starting point from which specificity can be engineered.
\n
\ud83d\udca1 Why This Is Cool
\n
This is the first de novo enzyme designed without borrowing from nature that matches natural catalytic efficiency. The rate enhancement of 3x10\u00b9\u00b3 is the highest for any de novo designed hydrolase. dEVA shows functional catalytic sites can emerge computationally from first principles.
\ud83d\udd2c High-affinity antibody variants often fail in production because of poor cellular expression. Researchers at Seoul National University combined AlphaFold3 and ProteinMPNN into a rescue workflow that identifies sequence-structure mismatches in CDR residues and corrects them, often with a single substitution, restoring expression while preserving binding affinity.
\n
\ud83e\uddec Affinity and expression are largely independent, meaning high-affinity, low-productivity (HALP) clones are not failures. They can be rescued.
\n
\u26a1 ProteinMPNN scores correlate strongly with cellular productivity across around 9,500 variants, revealing that expression is governed by how well CDR sequences fit their structural context, not just their biochemical properties.
Across 14 diverse HALP antibodies, single-residue substitutions restored expression in 11 cases, while preserving 80% or more of original binding affinity. This reframes failed candidates as recoverable assets rather than endpoints.
\n
2\ufe0f\u20e3 Sequence-Structure Compatibility as a Predictor
\n
ProteinMPNN scores serve as a reliable proxy for expression, providing a computationally cheap filter before expensive experimental validation.
\n
3\ufe0f\u20e3 Minimal Edits, Maximal Impact
\n
Rescue often required only one mutation, yielding up to 4-fold improvements in expression. These substitutions typically stabilise interactions within CDRs or between CDRs and the antibody framework.
\n
4\ufe0f\u20e3 Decoupling Affinity and Developability
\n
Because productivity and affinity landscapes are independent, expression can be improved without compromising binding, solving a long-standing trade-off in antibody engineering.
\n
\ud83d\udca1 Why This Is Cool
\n
Antibody design has long focused on finding better binders, but binding is only half the story. By reframing expression as a structural compatibility problem, failed candidates become fixable rather than disposable. This turns antibody engineering from a filtering step into a repair-and-optimise loop, expanding the usable therapeutic space.
\ud83d\udd2c Generative AI can propose drug candidates, but most fail at the same hurdle: they cannot be synthesised efficiently. Molecules that look promising on screen often require impractical chemistry to produce. SyntheMol-RL from Stanford and McMaster University uses reinforcement learning to navigate a chemical space of 46 billion synthesisable compounds, optimising for both antibacterial activity and aqueous solubility simultaneously.
\n
\ud83e\uddec Built on real chemical building blocks and validated reaction templates, the model generates molecules with guaranteed synthetic routes. This is not theoretical synthesisability. Every output comes with a concrete pathway from purchasable reagents.
\n
\u26a1 79 novel compounds were synthesised and tested. 13 showed potent in vitro activity against Staphylococcus aureus. Seven passed structural novelty filters against known antibiotics. One compound, synthecin, demonstrated efficacy in a murine wound infection model of methicillin-resistant S. aureus (MRSA).
Unlike virtual screening, SyntheMol-RL generates molecules with built-in synthesis plans, removing the bottleneck between computational hits and experimental validation.
\n
2\ufe0f\u20e3 Multi-Parameter Optimisation
\n
Reinforcement learning enables simultaneous tuning of activity, solubility, and synthesisability rather than optimising one property at the cost of others.
\n
3\ufe0f\u20e3 In Vivo Validation
\n
Synthecin\u2019s efficacy in a wound infection model moves AI-designed antibiotics beyond petri dish activity into preclinical relevance.
\n
4\ufe0f\u20e3 Generalisable Framework
\n
The architecture is target-agnostic. Swap the reward function and the same framework applies across therapeutic domains, not just antibiotics.
\n
\ud83d\udca1 Why This Is Cool
\n
Most generative models for drug discovery propose molecules that cannot be made, or can be made but do not work. SyntheMol-RL closes both gaps: it only proposes what chemistry can deliver, and it validated a compound through to animal models. Going from 46 billion possibilities to a single molecule treating MRSA-infected wounds in mice is the full loop from generative AI to preclinical candidate.
Building in BioAI is a monthly newsletter for those operating in, or interested in, the AI-enabled biology space. That\u2019s founders, technical leaders, and individual contributors working within areas like therapeutics, diagnostics, and tooling.
Joe\u2019s roundup centres on observations from within the space, including analysis of how teams are structuring themselves, what\u2019s changing in hiring, where funding is landing, what headlines mean for growth, and how BioAI companies are thinking about commercialising what they\u2019re building.
Each edition pulls from ongoing conversations with people doing the work day-to-day, as well as his own take on what\u2019s hit headlines that month.
Joe recruits in this space day-to-day, and so often speaks from that vantage point. He spends most of his time inside these teams, hiring for them, speaking with founders and senior talent across the market. The aim isn\u2019t to overstate where things are going, but to give a clear picture of what\u2019s actually happening and why it matters if you\u2019re hiring or looking to commercialise in BioAI.
The best competitions, hackathons, and community challenges in AI x life sciences, curated weekly. Know something worth featuring? Reply and let us know.
ELIXIR\u2019s annual international bioinformatics hackathon, running since 2018. 160+ participants, five days of collaborative coding on open bioinformatics infrastructure and tools. The call for project proposals opens March 16 and closes April 15 - so if you want to lead a project, that\u2019s your window.
\n
\n
Thanks for reading!
\n
\ud83d\udcac Get involved
\n
We\u2019re always looking to grow our community. If you\u2019d like to get involved, contribute ideas or share something you\u2019re building, fill out this form or reach out to me directly.
\n
Connect With Us
\n
Have questions or suggestions? We'd love to hear from you!
Thanks for reading Kiin Bio! Subscribe for free to receive new posts and support my work.
\n
\n\n\n
\n
\n","enclosure":{"link":"https://substack-post-media.s3.amazonaws.com/public/images/b58acda7-5298-448f-93e9-41f3a4328502_1200x630.png","type":"image/jpeg"},"categories":[]},{"title":"\ud83e\uddea Rowan: Computational Chemistry Without the Code","pubDate":"2026-04-28 17:01:35","link":"https://newsletter.kiin.bio/p/rowan-computational-chemistry-without","guid":"https://newsletter.kiin.bio/p/rowan-computational-chemistry-without","author":"Natasha Kilroy","thumbnail":"","description":"Deep Dive | Edition 17","content":"\n
Welcome back to the deep dive, where we break down the AI tools and data reshaping how new drugs are discovered. In each edition, we speak directly with the teams behind these tools to explain what they solve, how they work and where they are going next.
\n
\n
Keeping up with AI x life science news can get exhausting.
\n
It\u2019s scattered across LinkedIn, X, Substack, arXiv, Slack, newsletters... and you still somehow miss the things that actually matter. Too much noise, not enough signal.
\n
We\u2019re building something to fix that: a smarter, more powerful way to stay on top of what\u2019s actually relevant to you.
\n
But we want to build it with you, not just for you. Take 2 minutes to tell us what\u2019s missing. What you share will directly shape what we build, and you\u2019ll be the first to benefit from it.
Today we\u2019re looking at Rowan, a Boston-based startup of six that\u2019s building a web-based computational chemistry platform that lets medicinal chemists run advanced modeling workflows directly, without needing to manage scripts, infrastructure, or specialist software.. We spoke with co-founder Corin Wagen, an experimental organic chemist turned computational entrepreneur, about why the gap between medicinal chemists and computational tools has persisted for so long, and what it takes to close it.
\n
\n
\u201cThere are all these really smart, really talented chemists and scientists who are just not able to use computation to help them out. You always have to ask somebody else to do it. There\u2019s all these artificial barriers.\u201d
The tools exist. You can predict binding affinities, generate conformers, run molecular dynamics, dock compounds into protein structures. But actually using any of this typically requires programming expertise, command-line fluency, access to the right hardware, and the patience to stitch together a dozen different software packages that were never designed to work together.
\n
For most medicinal chemists, the people actually deciding which compounds to synthesise next, this means going through a computational chemist every time they want to run something. That handoff slows everything down. It creates bottlenecks, introduces miscommunication, and means that computation gets used selectively rather than routinely.
\n
The result: most drug discovery teams are making synthesis decisions with less computational insight than they could be: not because the science isn\u2019t there, but because the software gets in the way.
\n
\u201cYou should be a chemist to use Rowan, but you shouldn\u2019t need to be a programmer.\u201d
\n
\n
\ud83d\udca1 The Idea
\n
Rowan\u2019s answer is a web-based platform that organises computational chemistry into workflows, data in, data out, aligned to how scientists actually think about experiments.
\n
You have a compound and want to know how soluble it will be? That\u2019s a workflow. You have a binding pose and want to screen 100 analogues? That\u2019s a workflow. Under the hood it might be an ML model, a physics simulation, or a database lookup, but the scientist doesn\u2019t need to care about the plumbing.
\n
The founding team brings a deliberate mix: machine learning, software engineering (ex-Meta), quantum chemistry, product and business, and experimental organic chemistry. Wagen sees that breadth as essential. \u201cA lot of times the best ideas in chemistry come from people who\u2019ve journeyed outside chemistry and then bring back new ideas.\u201d
\n
The platform spans workflows from hit discovery to candidate selection, ligand-based methods (ML potentials, rapid quantum chemistry, conformers, reactivity, spectra) and, increasingly, structure-based drug discovery (docking, co-folding, molecular dynamics, and now free energy perturbation). Scientists pick what they need. Rowan doesn\u2019t prescribe a single workflow.
\n
Critically, Rowan is designed to fit into existing scientific software stacks. Teams can use it through the browser, through Python, or as part of agentic and automated pipelines, without replacing the tools they already rely on.
\n
\u201cThere\u2019s a lot more smart people outside Rowan than inside Rowan. We don\u2019t need to own the whole thing.\u201d
\n
\n
\u2699\ufe0f The FEP Release
\n
The headline addition is free energy perturbation (FEP), the gold-standard physics-based method for predicting how binding affinity changes across a series of related compounds. If you\u2019re optimising a lead and need to decide which of 100 analogues to actually synthesise, FEP tells you which ones are likely to bind better and which are duds, before you spend time and money making them.
\n
FEP has been around for decades, but two things have kept it out of mainstream medicinal chemistry workflows: it\u2019s expensive (historically around 10 GPU hours per compound) and it\u2019s complicated to set up and run.
\n
Rowan partnered with Forrest York and the open-source TMD engine (originally from Relay Therapeutics) to tackle both. The engineering improvements are dramatic: default settings run approximately 10 minutes per leg, with the potential to reach 1-2 minutes per leg with adjusted settings. That\u2019s a roughly 60x speedup over the literature standard.
\n
The speed comes from low-level optimisation and an algorithmic trick called local resampling, where instead of simulating the entire protein, the calculation focuses on the immediate neighbourhood around the ligand. \u201cIf you naively try to do that, it works very poorly,\u201d Wagen explains. \u201cBut it turns out if you very cleverly try to do that, it works very well.\u201d
\n
The end-to-end workflow can be run without any coding. Three steps: prepare your poses using analogue docking, build a perturbation graph showing which ligands to compare, and run the FEP calculations on cloud GPUs. Results stream back to your browser in real time.
The Rowan FEP workflow: users build a perturbation graph of ligand similarities and run binding affinity calculations directly in the browser, with results streaming back in real time.
\n
\n
\ud83d\udcca The Trade-Off
\n
The honest question: how accurate is it?
\n
Rowan\u2019s benchmarks show a mean absolute error of approximately 1.3 kcal/mol, compared to Schr\u00f6dinger\u2019s ~0.8 kcal/mol. The gap comes from two places: the speed-optimising approximations and the use of open force fields rather than Schr\u00f6dinger\u2019s proprietary ones.
\n
Wagen is transparent about this. \u201cOur benchmarks are a little bit worse than Schr\u00f6dinger\u2019s. They have amazing force fields. They\u2019re so good at force fields.\u201d
\n
But the metric that matters most in practice isn\u2019t absolute energy prediction, it\u2019s ranking. If FEP correctly tells you which compounds will bind better and which won\u2019t, you\u2019ve saved synthesis cycles regardless of whether the exact energy numbers are perfect. And on ranking, the performance gap narrows considerably.
\n
\u201cIf compounds are predicted to bind terribly, unless you really don\u2019t understand the binding mode, they\u2019re almost always bad binders. We have a lot of benchmark data showing this.\u201d
\n
The calculus is straightforward: if you can run FEP on 10\u2013100x more compounds than you can synthesise, at a fraction of the cost and time, slightly wider error bars are a trade-off most teams will take.
\n
Rowan publishes its benchmark data openly: you can explore the full results at benchmarks.rowansci.com.
Speed changes behaviour. At 10 GPU hours per compound, FEP is something you run occasionally on high-value decisions. At 10 minutes, it becomes routine, something you run on every idea before committing to synthesis. That shift from selective to systematic is the real unlock.
\n
No code, no setup. Most FEP implementations require protein preparation scripts, force field configuration, graph construction code, and hardware management. Rowan handles all of this in three browser-based workflows, while also providing API access for teams that want to automate and run FEP at scale.
\n
Open infrastructure, not a black box. Built on the open-source TMD engine and open force fields. Benchmarks are published. The platform is designed to complement existing tools via API, not replace entire workflows.
\n
Built for the medicinal chemist. Rowan is built to put computational insight directly in the hands of medicinal chemists, while still remaining useful to computational and platform teams.
\n
\n
\ud83d\udc8a Who It\u2019s For
\n
Rowan\u2019s primary customers are small-to-medium biotechs and pharma departments that don\u2019t have large internal computational tooling teams. The philosophy is to complement, not compete.
\n
\u201cWe often work with companies where they\u2019re small enough that they don\u2019t have anybody who they can ask to do the work that we do for them,\u201d Wagen says.
\n
The platform recently passed 10,000 users and has generated over 40 publications. Pricing for FEP runs at approximately $5-10 per edge for platform customers, or $25 per ligand through a managed fee-for-service option where a Rowan scientist handles the analysis.
\n
\n
\ud83d\udd2e The Future
\n
The near-term goal is integration into fast-moving drug discovery programmes through active pilots. The dream: automated nightly runs from SMILES strings to predicted binding affinities, with a digest landing in your inbox each morning showing which AI-generated analogues are worth pursuing.
\n
\u201cI suspect that the dream is that we can just blindly put SMILES in, go all the way to binding affinities, and that runs every single night.\u201d
\n
That dream is not yet reality. Fully automated structure-based modeling is still not trivial, particularly in pose preparation and graph construction, and Wagen is candid that the final stretch depends on solving messy, program-specific edge cases and learning from their early FEP customers. Beyond automation, the roadmap includes custom force-field fitting for client compound series (to close the accuracy gap),, pre-FEP triage tools, and continued speed optimisation targeting under one minute per compound for large libraries.
\n
\u201cI hope this doesn\u2019t take us a few years. I hope this can happen in 2026 for Rowan.
We\u2019re always looking to grow our community. If you\u2019d like to get involved, contribute ideas or share something you\u2019re building, fill out this form or reach out to me directly.
Welcome back to your weekly dose of AI news for Life Science!
\n
\n
Keeping up with AI x life science news can get exhausting.
\n
It\u2019s scattered across LinkedIn, X, Substack, arXiv, Slack, newsletters... and you still somehow miss the things that actually matter. Too much noise, not enough signal.
\n
We\u2019re building something to fix that: a smarter, more powerful way to stay on top of what\u2019s actually relevant to you.
\n
But we want to build it with you, not just for you. Take 2 minutes to tell us what\u2019s missing. What you share will directly shape what we build, and you\u2019ll be the first to benefit from it.
\ud83d\udd2c Spatial transcriptomics captures gene expression within tissue architecture, but existing models either ignore spatial relationships or flatten each cell into a simple feature vector. They miss the interplay between a cell\u2019s internal gene programmes and its tissue neighbourhood.
\n
HEIST from Yale models tissues as hierarchical graphs. The upper level captures spatial relationships between cells, while each cell is represented by its own gene co-expression network. Cross-level message passing connects the two, letting internal regulation and spatial context inform each other.
\n
\ud83e\uddec Pretrained on 22.3 million cells from 124 tissues across 15 organs using spatially-aware contrastive learning and masked autoencoding. HEIST uses flexible gene vocabularies rather than fixed gene sets, so it generalises to unseen genes and even spatial proteomics without retraining.
\n
\u26a1 Unsupervised analysis reveals spatially informed cell subpopulations missed by prior models. Downstream, HEIST achieves state-of-the-art performance in clinical outcome prediction, cell type annotation, and gene imputation across multiple spatial technologies.
Transfers from transcriptomics to proteomics without retraining, making it applicable across spatial profiling technologies.
\n
2\ufe0f\u20e3 Tissue Microenvironment Discovery
\n
The hierarchical design captures spatially defined subpopulations that flat models miss, enabling more nuanced tissue phenotyping.
\n
3\ufe0f\u20e3 Clinical Outcome Prediction
\n
Patient-level embeddings from spatial data support tasks like treatment response and survival prediction.
\n
4\ufe0f\u20e3 Flexible Gene Vocabularies
\n
By avoiding fixed gene sets, HEIST handles new panels and custom targets without architectural changes.
\n
\ud83d\udca1 Why This Is Cool
\n
Most spatial models look at where cells are or what they express. HEIST does both through hierarchical graph modelling. Generalising to proteomics without retraining suggests these representations capture something fundamental about how cells organise within tissues.
\ud83d\udd2c ML models for antibody design are held back by a data bottleneck: not enough structural training data linking designed antibodies to validated binding outcomes. Existing datasets are small, biased towards natural antibodies, and lack systematic off-target measurements.
\n
A-Alpha Bio built SEPIA (Synthetic Epitope Atlas), pairing over 26 million on-target and off-target binding measurements with computationally designed VHH-antigen structures. Using their AlphaSeq yeast-based platform, they measured binding affinities and specificity across thousands of de novo synthetic epitope proteins designed to bind VHH nanobodies.
\n
\ud83e\uddec Each designed VHH-SEP pair comes with both structural predictions and experimental binding data, so models can learn what makes a designed complex actually bind versus what only looks good computationally.
\n
\u26a1 Across thousands of variants, SEPIA validates strong specificity and provides the negative data most antibody datasets lack. Positive and negative measurements at this scale give ML models a clearer signal for learning specificity, not just affinity.
26 million measurements paired with designed structures create a purpose-built resource for next-generation antibody design models.
\n
2\ufe0f\u20e3 Specificity, Not Just Affinity
\n
Systematic off-target measurements let models learn what not to bind, addressing a major blind spot in current datasets.
\n
3\ufe0f\u20e3 Closing the Design-Validation Loop
\n
Linking computational designs directly to high-throughput experimental readouts enables rapid iteration on antibody engineering.
\n
4\ufe0f\u20e3 Nanobody-Focused Design
\n
VHH nanobodies are increasingly important as therapeutics. A large-scale, VHH-specific dataset accelerates this growing field.
\n
\ud83d\udca1 Why This Is Cool
\n
The gap between computational antibody design and experimental reality has always been the data. SEPIA fills it with 26 million purpose-built binding measurements, including both what works and what does not. Models trained on real specificity data at this scale can finally learn to design antibodies that are specific, not just tight binders.
\ud83d\udd2c Modern hospitals generate vast multimodal data across labs, imaging, notes, medications, and procedures, but it sits in disconnected systems. No existing model integrates the full breadth and temporal depth of a clinical record into one unified representation.
\n
Apollo from Harvard Medical School does exactly that. Trained on over 30 years of longitudinal records from Mass General Brigham, it unifies 28 modalities and 12 major specialties into a shared embedding space, building an \u201catlas of medical concepts.\u201d
\n
\ud83e\uddec Apollo processes entire care journeys as sequences of structured and unstructured events, compressing them into virtual patient representations. Its vocabulary spans over 100,000 unique medical events alongside clinical images and free-text notes. Feature attribution confirms predictions align with clinically interpretable biomarkers.
\n
\u26a1 Evaluated across 322 tasks on 1.4 million held-out patients: disease onset prediction up to five years ahead (95 tasks), disease progression (78), treatment response (59), adverse event risk (17), and hospital operations (12). Apollo also functions as a multimodal medical search engine across 61 retrieval tasks.
Predicting disease onset years in advance from the full patient record enables proactive intervention rather than reactive care.
\n
2\ufe0f\u20e3 Treatment Response Prediction
\n
Drawing on a patient\u2019s complete multimodal history supports more personalised therapy decisions across specialties.
\n
3\ufe0f\u20e3 Multimodal Medical Search
\n
Text and image queries against patient embeddings create a clinical search engine for cohort identification and case matching.
\n
4\ufe0f\u20e3 Interpretable Predictions
\n
Feature attribution shows outputs align with known biomarkers, bridging AI predictions and clinical reasoning.
\n
\ud83d\udca1 Why This Is Cool
\n
This is the first model to compress decades of multimodal clinical data into unified patient embeddings at hospital system scale. Moving from narrow, task-specific clinical models to holistic representations that predict disease, treatment response, and adverse events from one embedding is a fundamental shift for clinical AI.
The best competitions, hackathons, and community challenges in AI x life sciences, curated weekly. Know something worth featuring? Reply and let us know.
\n
\ud83d\udcc4Recap post: BIOMICS Hackathon | Feb 23-25
\n
The BIOMICS hackathon at EMBL-EBI brought together computational biologists and software engineers from Portugal, Spain, Germany, and the UK for three days of building. Five challenge tracks covered everything from statistical tools to building software for a brand new microscopy technique from scratch.
\n
Every team built something visual, reflecting a shift away from command-line-only workflows. This was also one of the first hackathons where AI coding agents like Claude Code were widely used across teams, and the difference in what could be achieved in three days was significant. One participant described these tools as an \u201cexoskeleton\u201d that amplifies existing ability.
\n
BIOMICS is a EU Horizon-funded project twinning GIMM Lisbon with EMBL-EBI, CRG Barcelona, and ETH Zurich to strengthen biomedical data science training and collaboration. More events are planned throughout the year.
ELIXIR\u2019s annual international bioinformatics hackathon, running since 2018. 160+ participants, five days of collaborative coding on open bioinformatics infrastructure and tools. The call for project proposals opens March 16 and closes April 15 - so if you want to lead a project, that\u2019s your window.
\n
\n
Thanks for reading!
\n
\ud83d\udcac Get involved
\n
We\u2019re always looking to grow our community. If you\u2019d like to get involved, contribute ideas or share something you\u2019re building, fill out this form or reach out to me directly.
\n
Connect With Us
\n
Have questions or suggestions? We'd love to hear from you!
Thanks for reading Kiin Bio! Subscribe for free to receive new posts and support my work.
\n
\n\n\n
\n
\n","enclosure":{"link":"https://substack-post-media.s3.amazonaws.com/public/images/b088ff9f-9241-4bf1-a13f-98085f19924a_1200x630.png","type":"image/jpeg"},"categories":[]},{"title":"\ud83e\udd7c Luvida: Bringing the Whole of a Patient\u2019s Life Into Clinical Trials","pubDate":"2026-04-21 17:02:20","link":"https://newsletter.kiin.bio/p/luvida-bringing-the-whole-of-a-patients","guid":"https://newsletter.kiin.bio/p/luvida-bringing-the-whole-of-a-patients","author":"Natasha Kilroy","thumbnail":"","description":"Deep Dive | Edition 16","content":"\n
Welcome back to the deep dive, where we break down the AI tools and data reshaping how new drugs are discovered. In each edition, we speak directly with the teams behind these tools to explain what they solve, how they work and where they are going next.
\n
\n
Keeping up with AI x life science news can get exhausting.
\n
It\u2019s scattered across LinkedIn, X, Substack, arXiv, Slack, newsletters... and you still somehow miss the things that actually matter. Too much noise, not enough signal.
\n
We\u2019re building something to fix that: a smarter, more powerful way to stay on top of what\u2019s actually relevant to you.
\n
But we want to build it with you, not just for you. Take 2 minutes to tell us what\u2019s missing. What you share will directly shape what we build, and you\u2019ll be the first to benefit from it.
The failure rate in clinical development is well known but no less striking for it. Around 90% of clinical trials still fail. And buried inside that number is a problem the industry has not yet confronted: trials are still being designed on incomplete patient information, slowing recruitment, driving up attrition, and causing costly delays, ultimately putting medicines out of reach for the patients who need them most.
The industry spends an eyewatering $400 billion a year on that failure. And while the causes are multiple, a significant portion of that failure is traceable to decisions made at the protocol design stage, before a single patient is enrolled.
\n
Trial teams are making high-stakes decisions about patient populations, eligibility criteria, endpoints, and recruitment strategies on a partial evidence base. Biology is well represented. But the patient\u2019s life is not. The result is a systematic underestimation of the factors that actually determine recruitment and screening success, patient dropout, adherence, and the need for subsequent protocol amendments.
We sat down with co-founders Hannah Amies and James Malone to understand what they\u2019re working on. Hannah\u2019s background spans biomedicine, consulting, head of product at BenevolentAI and epidemiology at Oxford. James brings computer science, bioinformatics, and a career spanning the European Bioinformatics Institute, his own acquired data curation company, and CTO roles at SciBite and Benevolent AI. Between them they cover both sides of the problem: the data science and the clinical domain knowledge.
\n
\n
\ud83d\udd34 The Problem
\n
Clinical trial protocol design remains a surprisingly manual, consensus driven process for a field built on evidence. Getting a protocol ready involves assembling internal teams, bringing in external key opinion leaders, iterating over months, and more excel spreadsheets than are possible to manage. It can take up to 18 months, just for the design phase. And it\u2019s not especially data-driven. \u201cIt\u2019s very expert opinion driven,\u201d James told us. \u201cThat can be very advantageous, you need that expertise. But it does mean biases creep in. Evidence is scattered across documents such as historical protocols, published literature, amendment documents, recruitment and on-trial data, and regulatory feedback. This is also a data problem.\u201d
\n
The result? Around 50% of trials end up requiring protocol amendments, averaging 3.3 per trial. Each one costs roughly $500,000 and burns at least three months waiting for regulatory sign-off. Do the maths across multiple trials per drug and you\u2019re looking at hundreds of millions in lost on-patent revenue, and years of delay before a drug reaches a patient who needs it. \u201cMost people working in the space are just doing it because they believe in getting good drugs into the right patients\u2019 hands,\u201d James said. That\u2019s the real cost of a broken process.
\n
\n
\ud83d\udcca The Missing Data
\n
The core issue is an incomplete patient picture. Clinical and biomedical data captures areas like biology, genotypic profiles, disease characteristics and prior treatment history. What it does not capture is the much broader set of variables that determine whether a given patient is recruited, adheres to treatment, or withdraws early.
\n
The data exists, it is simply not being used. Hannah\u2019s path to founding Luvida started not in a lab, but in Liverpool, implementing electronic patient records across three hospitals. That\u2019s where she first noticed the gap: mountains of health data, almost entirely underleveraged. Epidemiology at Oxford sharpened the picture. \u201cA lot of this stuff we have evidence for\u201d she said, \u201cbut a lot of it is buried in research papers and not being leveraged at scale\u201d
\n
Luvida\u2019s answer is what they call Electronic Life Records, a proprietary data layer that builds a more complete picture of the patient than clinical and biomedical data alone. It is that richer picture that changes what you can predict, and how accurately.
\n
\n
\ud83d\udca1 The Idea: Expert in the Loop, Not AI in Charge
\n
Luvida\u2019s platform isn\u2019t trying to replace clinical operations teams. James was clear: \u201cWe don\u2019t want to come in and look like we\u2019re replacing a clinical trial team of medical writers.\u201d The goal is to speed up the parts of the job that involve synthesising signals from disparate, messy data sources, then hand that signal back to people who know what to do with it.
\n
The platform works within a familiar authoring environment, think Google Docs-style interface, where AI-driven suggestions surface like comments: accept, reject, or edit. Every recommendation is tied directly to the evidence for review, which is used for decision-making and regulatory justification. There\u2019s also a chat interface for question-answering. But the core of what Luvida does is pull together clinical and biomedical data, curated historical trial data, and patient lifestyle and behavioural data to flag where a trial is likely to fail before it starts.
Trial Explorer: Trial explorer enables access to Luvida\u2019s curated trial and publication datasets, allowing you to deeply understand and interrogate the trial and literature landscape and gain analytical insights for trials across therapeutic areas.
\n
Questions like: which populations are most likely to drop out? Where will recruitment stall? Are your eligibility criteria accidentally excluding a particular ethnic group? Does this background population have co-morbidities you haven\u2019t accounted for? Is the evidence strong enough for regulatory review?
\n
Crucially, it pairs AI pattern-finding with strict rule-based approaches. \u201cAI is great at looking for big patterns,\u201d James said, \u201cbut many modern models are trained to be helpful, which means they can sometimes infer things that aren\u2019t really there.\u201d So Luvida enforces rigour: everything is auditable, everything is traceable, and recommendations are evidence-based.
\n
\n
\ud83d\udce2 Why It\u2019s Different
\n
For large pharma, Luvida gives teams the evidence base to navigate internal governance faster. Instead of one expert\u2019s opinion against another\u2019s, teams walk into meetings with data. For smaller biotechs who rely entirely on CROs and often feel, as Hannah put it, \u201cpretty disempowered\u201d, it\u2019s even more significant. Luvida arms them with evidence to push back and ask better questions.
\n
For rare diseases, where historical trial data is thin by definition, the platform can identify analogous disease areas and trials that work in adjacent spaces. \u201cThat\u2019s something that AI is really good at identifying,\u201d Hannah said. \u201cThe patterns. And that\u2019s really where our models come into play.\u201d
Luvida\u2019s proprietary models: Luvida predicts risks, mitigations and success factors, all backed by evidence and regulatory guidance to enable you to de-risk your protocols.
\n
The early validation numbers are striking. When Luvida ran their models against past trials, feeding in initial protocols while blinding the system to the amendments that followed, it predicted 40- 50% of those subsequent amendments. In a prototype. The cost of each predicted amendment that doesn\u2019t happen? Hundreds of thousands of dollars and three months of time.
\n
\n
\ud83d\udd2e The Future
\n
Luvida is approximately a year old and is working with forward-thinking customers on live trial design programmes. Data is processed by indication to ensure data quality, a deliberate intention that keeps model outputs high quality and trustworthy. The platform already covers 500K+ trials, with 200K+ enriched with publications. From this, for example, we\u2019ve already surfaced 87K+ known adherence risks.
\n
The roadmap includes key opinion leaders contributing directly to the platform to streamline stakeholder management, and eventually patients too, to ensure the patient voice is brought in from the TPP, with protocols translated into plain language. Scenario modelling alternative trial designs enables pressure-testing of real-world on-trial risks before a trial even starts.
\n
There\u2019s also an API, so customers can plug Luvida into their own internal tooling and models, rather than adding another standalone system to an already crowded stack.
\n
The platform is also designed with an eye on where regulation is heading, FDA diversity action plans and the NHS 10-year plan both point in the same direction: trials that are designed to reflect the real-world diversity of patient populations. Luvida\u2019s richer patient picture makes that not just possible, but built in from the start.
\n
The incomplete patient picture has been one of the industry\u2019s most persistent and expensive blind spots for decades. Luvida is building the infrastructure to close it.
Luvida is headquartered in London and already partnering with CROs, pharma, and biotech across the EU and US. For a limited time, they\u2019re offering exclusive value assessments. To see Luvida in action on one of your protocols, secure your spot at enquiries@luvida.co.uk.
James Malone, CTO and co-founder and Hannah Amies, CEO and founder.
\n
\n
Thanks for reading Kiin Bio Weekly!
\n
\ud83d\udcac Get involved
\n
We\u2019re always looking to grow our community. If you\u2019d like to get involved, contribute ideas or share something you\u2019re building, fill out this form or reach out to me directly.
Welcome back to your weekly dose of AI news for Life Science!
\n
\n
Keeping up with AI x life science news can get exhausting.
\n
It\u2019s scattered across LinkedIn, X, Substack, arXiv, Slack, newsletters... and you still somehow miss the things that actually matter. Too much noise, not enough signal.
\n
We\u2019re building something to fix that: a smarter, more powerful way to stay on top of what\u2019s actually relevant to you.
\n
But we want to build it with you, not just for you. Take 2 minutes to tell us what\u2019s missing. What you share will directly shape what we build, and you\u2019ll be the first to benefit from it.
\ud83c\uddfa\ud83c\uddf8 Meet Us at Bio-IT, Boston
\n
\ud83e\udd1d Our CEO Filippo and CTO Bogdan will be at Bio-IT World in Boston. If you are attending and would love to connect in person, reach out and let us know.
\ud83d\udd2c Predicting protein melting temperatures (Tm) is critical for enzyme engineering: thermostable enzymes last longer and tolerate harsher conditions. But most AI predictors train on mass-spectrometry data from whole cells, which is fundamentally different from purified protein measurements. The Loschmidt Labs team found near-zero correlation (r = 0.05) between proteomics and biophysical Tm datasets for the same proteins.
\n
TmProt 1.0 from Loschmidt Laboratories (Masaryk University) fixes this by training exclusively on biophysically measured melting temperatures. Built on ESM-2 fine-tuned with LoRA, it is now live on Hugging Face.
\n
\ud83e\uddec The team assembled ProMelt, a curated set of 45,441 proteins with experimental Tm values, validated across 5 independent biophysics test sets. TmProt outperforms TemBERTure, DeepSTABp, and SaProt across most benchmarks.
\n
\u26a1 Absolute Tm regression remains hard, but TmProt excels where it matters most: ranking thermostable candidates. For classifying proteins with Tm at or above 60 degrees C, it achieves an AUC of 0.75, and is fully integrated into EnzymeMiner 2.0 for end-to-end enzyme mining.
Best used as a ranking tool to prioritise thermostable candidates from large sequence sets, making it practical for early-stage enzyme discovery campaigns.
\n
2\ufe0f\u20e3 Training Data Matters More Than Architecture
\n
The near-zero correlation between proteomics and biophysical Tm explains why previous predictors underperformed. Curating ProMelt was the single biggest driver of improvement.
\n
3\ufe0f\u20e3 Lightweight and Accessible
\n
LoRA fine-tuning keeps the model efficient enough to run as a free web server, lowering the barrier for labs without GPU infrastructure.
\n
4\ufe0f\u20e3 Integrated Discovery Pipeline
\n
Full integration with EnzymeMiner 2.0 means researchers go from sequence database to stability-ranked shortlist in one workflow.
\n
\ud83d\udca1 Why This Is Cool
\n
Data quality beating model complexity. The proteomics vs biophysical Tm disconnect explains years of underwhelming predictions. By fixing the data rather than stacking layers, TmProt delivers where enzyme engineers actually need it.
\ud83d\udd2c Proteins undergo conformational changes upon ligand binding, yet most deep learning generative models treat pockets as rigid, generating molecules for a single frozen conformation. The pocket shape in a crystal structure may not be the one your molecule actually binds.
\n
YuelDesign from the Dokholyan lab at the University of Virginia jointly models both the pocket structure and ligand conformation, allowing protein and molecule to co-adapt during generation.
\n
\ud83e\uddec Two diffusion processes run simultaneously: an elucidated diffusion model (EDM) for 3D coordinates and a discrete denoising diffusion model (D3PM) for atom types. Both use E3former to maintain rotational and translational equivariance.
\n
\u26a1 The result is molecules with favourable drug-likeness, low synthetic complexity, diverse functional groups, and docking energies comparable to native ligands. By letting the pocket breathe during generation, YuelDesign captures induced-fit effects that rigid methods miss.
Jointly diffusing pocket and ligand explores conformational states that only emerge upon binding, capturing selectivity-driving dynamics rigid models cannot access.
\n
2\ufe0f\u20e3 Synthesisable and Drug-Like Output
\n
Generated molecules score well on synthetic accessibility out of the box, reducing the gap between computational hits and what a medicinal chemist would actually make.
\n
3\ufe0f\u20e3 Diverse Chemical Exploration
\n
D3PM atom-type diffusion produces varied functional groups rather than collapsing to a narrow series, giving broader chemical space coverage per run.
\n
4\ufe0f\u20e3 Beyond Rigid Docking
\n
Treating flexibility explicitly moves generative design closer to real protein-ligand recognition, where both partners adjust shape upon binding.
\n
\ud83d\udca1 Why This Is Cool
\n
The rigid-pocket assumption has limited structure-based design for decades. YuelDesign tackles it with a clean dual-diffusion architecture that still produces drug-like, synthetically accessible molecules. A meaningful step for generative molecular design.
\ud83d\udd2c Traditional enzyme design starts with a theozyme: a hand-crafted catalytic geometry based on detailed mechanistic knowledge. This works when you understand the mechanism, but limits you to chemistries where that knowledge exists.
\n
DISCO (Diffusion for Sequence-Structure Co-design) from Caltech co-generates sequence and structure simultaneously using diffusion over discrete amino acids and continuous 3D coordinates, conditioned only on DFT-derived reactive intermediates. No predefined catalytic motifs needed.
\n
\ud83e\uddec Given a target chemistry, DISCO explores catalytic solutions freely, generating novel active sites and repurposing unrelated protein folds for catalysis. The model jointly optimises sequence and structure in a single pass, preserving their natural interdependence.
\n
\u26a1 On the STUDIO-179 benchmark, DISCO generated the highest proportion of co-designable protein-ligand complexes for 178 of 179 cases. High-performing enzymes were found from just 90 designs, and for C(sp3)-H insertion, a poorly understood reaction, DISCO-designed enzymes matched variants from extensive directed evolution.
Conditioning on reactive intermediates rather than predefined geometries lets DISCO tackle chemistries where mechanistic knowledge is incomplete or unavailable.
\n
2\ufe0f\u20e3 Evolvable by Design
\n
Generated enzymes respond well to directed evolution, meaning computational design and experimental optimisation work as complementary stages.
\n
3\ufe0f\u20e3 Novel Folds for Novel Chemistry
\n
DISCO repurposes unrelated protein scaffolds and generates new active sites, expanding the design space beyond known enzyme families.
\n
4\ufe0f\u20e3 Reduced Experimental Screening
\n
Functional enzymes from just 90 designs dramatically reduces screening burden compared to traditional library approaches.
\n
\ud83d\udca1 Why This Is Cool
\n
Enzyme design has always been bottlenecked by mechanistic knowledge. DISCO removes that constraint. Matching directed evolution on a poorly understood reaction from de novo designs suggests AI can now explore catalytic solutions humans would not think to try.
The best competitions, hackathons, and community challenges in AI x life sciences, curated weekly. Know something worth featuring? Reply and let us know.
A free, two-session workshop co-hosted by Manuel Corpas and Segun Fatumo, running entirely in Google Colab. No setup, no cost, just a browser. Session one covers variant interpretation using Ensembl VEP, ClinVar, and ACMG criteria. Session two runs a full GWAS pipeline with ClawBio, including polygenic risk scores and locus fine-mapping. Built to remove barriers for researchers anywhere in the world.
ELIXIR\u2019s annual international bioinformatics hackathon, running since 2018. 160+ participants, five days of collaborative coding on open bioinformatics infrastructure and tools. The call for project proposals opens March 16 and closes April 15 - so if you want to lead a project, that\u2019s your window.
\n
\n
Thanks for reading!
\n
\ud83d\udcac Get involved
\n
We\u2019re always looking to grow our community. If you\u2019d like to get involved, contribute ideas or share something you\u2019re building, fill out this form or reach out to me directly.
\n
Connect With Us
\n
Have questions or suggestions? We'd love to hear from you!
Thanks for reading Kiin Bio! Subscribe for free to receive new posts and support my work.
\n
\n\n\n
\n
\n","enclosure":{"link":"https://substack-post-media.s3.amazonaws.com/public/images/f79013e0-ba6a-4df2-9791-1a6892146835_1200x630.png","type":"image/jpeg"},"categories":[]},{"title":"\ud83c\udfe5 Our Future Health: Building the World's Largest Health Research Programme","pubDate":"2026-04-14 17:01:47","link":"https://newsletter.kiin.bio/p/our-future-health-building-the-worlds","guid":"https://newsletter.kiin.bio/p/our-future-health-building-the-worlds","author":"Natasha Kilroy","thumbnail":"","description":"Deep Dive | Edition 15","content":"\n
Welcome back to the deep dive, where we break down the AI tools and data reshaping how new drugs are discovered. In each edition, we speak directly with the teams behind these tools to explain what they solve, how they work and where they are going next.
\n
\n
Keeping up with AI x life science news can get exhausting.
\n
It\u2019s scattered across LinkedIn, X, Substack, arXiv, Slack, newsletters... and you still somehow miss the things that actually matter. Too much noise, not enough signal.
\n
We\u2019re building something to fix that: a smarter, more powerful way to stay on top of what\u2019s actually relevant to you.
\n
But we want to build it with you, not just for you. Take 2 minutes to tell us what\u2019s missing. What you share will directly shape what we build, and you\u2019ll be the first to benefit from it.
Today we\u2019re having a look at Our Future Health, the UK\u2019s largest health research programme, building a five\u2011million\u2011strong cohort to transform prevention, early detection and treatment. To understand how Our Future Health balances participant experience, scientific discovery, and data security, we sat down with James Siddle, a digital health consultant focusing on health insights within the programme.
Large-scale health research programmes have shown how powerful population-level data can be for advancing prevention, diagnosis, and treatment. By linking biological samples, lifestyle information, and medical records, these initiatives have helped uncover genetic risk factors, guide new therapies, and inform public health policy.
\n
Yet as the ambition and scale of such programmes grow, so do the challenges. Collecting and safeguarding highly sensitive data from millions of people demands exceptional security and governance. At the same time, public expectations are evolving; participants increasingly want to understand how their contribution supports discovery and how it might help their own health.
\n
The central question for modern population health research is therefore how to build a system that serves both sides: one that enables cutting-edge science while ensuring every volunteer gains something valuable and personally meaningful in return.
\n
\n
\ud83d\udca1 The Idea\n
\n
Our Future Health brings together up to five million volunteers across the UK to help develop new ways to prevent, detect and treat diseases. Each participant contributes to creating one of the most detailed pictures of population health ever assembled, a resource designed to represent the full diversity of the UK and to power discoveries that improve everyone\u2019s health.
\n
Taking part involves completing health and lifestyle questionnaires, attending a short clinic appointment to have physical measurements taken, and providing a small blood sample. This information is securely linked with participants\u2019 NHS records to build a rich, de-identified dataset that reflects real-world health across age, ethnicity, and region.
\n
By combining this scale and depth of data with a focus on participant experience, Our Future Health aims to accelerate understanding of how different factors, such as genetics, environment, and behaviour interact to influence disease. Researchers from academia, the NHS, charities and industry can apply to access the data within a secure Trusted Research Environment, enabling them to explore new questions and develop better ways to predict, prevent and treat illness.
\n
At its core, the programme is designed to create a virtuous cycle between public participation and research impact: people contribute information that helps science move forward, and the knowledge gained feeds back into better health outcomes for future generations.
Our Future Health is different to other health research programmes. Its design choices make it a discovery engine with unique advantages:
\n
\n
Participant-first interface Our Future Health is developing an experience that prioritises participant clarity, ensuring that information is presented in a way that feels useful and easy to understand.
\n
Built-in recontact The programme plans to enable participants to be re-invited to future studies, creating opportunities for more dynamic research collaboration. As James noted, recontact requires \u201ccareful communication so people don\u2019t feel alarmed,\u201d a principle that guides every design decision.
\n
Scale & diversity With over two and a half million participants already enrolled, Our Future Health is the world\u2019s largest health research programme of its kind and is on track to become the most diverse health cohort worldwide. Diversity across age, ethnicity, and geography ensures that the discoveries we make can benefit everyone.
\n
Security by design James emphasised that \u201cinformation security is the top priority.\u201d Our Future Health adheres to ISO 27001 standards and works closely with the UK\u2019s National Cyber Security Centre. These measures are vital to maintaining participant trust and ensuring long-term sustainability.
\n
Active research underway Our Future Health is already enabling discovery. In June 2025, the first peer-reviewed study using its data was published in BMJ Mental Health by researchers at the University of Edinburgh. Analysing information from over 1.5 million participants, the team found that people with chronic inflammatory conditions may face nearly double the risk of mental-health issues compared with others.
\n
\n
This landmark paper marks the first of many studies set to use Our Future Health data to advance understanding of disease prevention, detection and treatment across a wide range of conditions.
As Our Future Health continues to grow toward its five-million-participant goal, its research value will expand exponentially. A recontactable, diverse population of this scale enables discoveries that could:
\n
\n
Identify early markers of disease through long-term follow-up.
\n
Support targeted prevention strategies by highlighting who might benefit most from early interventions.
\n
Accelerate clinical trials by improving the speed and representativeness of recruitment.
\n
\n
For James and the Our Future Health team, success means giving participants something meaningful in return while advancing discoveries that improve health outcomes for everyone.
\n
\ud83d\udc68\u200d\ud83d\udcbb Get in touch with James.
We\u2019re always looking to grow our community. If you\u2019d like to get involved, contribute ideas or share something you\u2019re building, fill out this form or reach out to me directly.
Welcome back to your weekly dose of AI news for Life Science!
\n
\n
Keeping up with AI x life science news can get exhausting.
\n
It\u2019s scattered across LinkedIn, X, Substack, arXiv, Slack, newsletters... and you still somehow miss the things that actually matter. Too much noise, not enough signal.
\n
We\u2019re building something to fix that: a smarter, more powerful way to stay on top of what\u2019s actually relevant to you.
\n
But we want to build it with you, not just for you. Take 2 minutes to tell us what\u2019s missing. What you share will directly shape what we build, and you\u2019ll be the first to benefit from it.
\ud83d\udd2c Protein embeddings power everything from interaction prediction to functional annotation. But how we evaluate them matters just as much as how we build them. Random train/test splits let models cheat by memorizing well-connected proteins, inflating performance on benchmarks while hiding failures on the long tail of understudied proteins.
\n
GATSBI (Graph Attention with Split-Boosted Inference) is a framework from Stanford that builds context-aware protein embeddings by integrating heterogeneous biological data, including PPIs from STRING, co-expression patterns, tissue-specific functional associations from HumanBase, and ESM-2 sequence representations, into a unified graph attention network.
\n
\ud83e\uddec The network covers 18,049 human proteins with 1.5M+ edges across three relationship types. Graph attention layers learn to weight different edge types and tissue contexts during message passing, producing embeddings that capture both local interactions and broader functional context.
\n
\u26a1 The key insight is evaluation design. GATSBI introduces biologically motivated splits: edge splits (testing unseen relationships between known proteins) and node splits (testing entirely unseen proteins with less than 30% sequence identity). Across interaction prediction, function classification, and functional set prediction, GATSBI consistently outperforms PINNACLE, with the largest gains for understudied proteins (AUROC +0.259, AUPRC +0.290 on functional sets).
1\ufe0f\u20e3 Plug-and-Play Protein Representations
\n
Embeddings are pretrained and downloadable, making them drop-in features for downstream models predicting interactions, function, or pathway membership without retraining the graph.
\n
2\ufe0f\u20e3 Honest Benchmarking for Understudied Proteins
\n
The node split reveals how models actually perform on proteins with limited prior evidence, exposing a critical gap that random splits conceal in current benchmarking.
High-confidence predictions flagged as errors turned out to reflect real but unannotated relationships, suggesting the model captures genuine biology beyond current annotations.
\n
4\ufe0f\u20e3 Boosting the Long Tail of the Proteome
\n
The heterogeneous graph adds the most information for low-degree proteins, precisely the understudied targets where computational predictions matter most for discovery.
\n
\ud83d\udca1 Why This Is Cool
\n
The lesson here goes beyond protein embeddings. How you split your data determines what your benchmark actually measures. GATSBI shows that when you evaluate properly, the gap between methods changes dramatically, and models that look equivalent on well-studied proteins diverge sharply on the understudied ones that matter for real discovery.
\ud83d\udd2c Current imaging technologies can visualize tens of proteins simultaneously. A single human cell contains thousands. This gap means we still lack a complete picture of how the proteome is spatially organized, and brute-force experimental coverage is not practical at this scale.
\n
ProtiCelli is a deep generative model from Stanford and KTH, trained on 1.23 million immunofluorescence images from the Human Protein Atlas, that simulates microscopy images for 12,800 human proteins using just three cellular landmark stains as input (nucleus, microtubules, ER).
\n
\ud83e\uddec Given only the landmark channels, the model predicts what the protein channel would look like, effectively hallucinating biologically accurate fluorescence patterns. It generalizes to unseen cell types and drug perturbations, preserves hierarchical subcellular organization, and even recapitulates known protein-protein interactions from the spatial patterns alone.
\n
\u26a1 The team generated Proteome2Cell: 30.7 million simulated images representing 2,400 virtual cells across 12 human cell lines, now integrated into the Human Protein Atlas. The model can also infer drug-induced changes in protein localization from cell morphology alone, without ever seeing the drug-treated protein images during training.
Enables imaging experiments that would take years and millions of dollars to run physically, compressing exploration of protein spatial organization into a single model inference.
\n
2\ufe0f\u20e3 Drug Mechanism-of-Action Discovery
\n
Predicts how compounds alter protein localization from cell morphology alone, without staining for every target, potentially accelerating early-stage perturbation screening.
\n
3\ufe0f\u20e3 Orthogonal Interaction Signal
\n
Predicted co-localization patterns correlate with known protein-protein interactions, providing a spatial signal that complements sequence-based interaction predictions.
\n
4\ufe0f\u20e3 A Public Resource for the Community
\n
Proteome2Cell (30.7M images across 12 cell lines) is freely available through the Human Protein Atlas, giving researchers a shared baseline for cell biology and drug discovery.
\n
\ud83d\udca1 Why This Is Cool
\n
This is the microscopy equivalent of protein structure prediction. Just as AlphaFold gave us predicted structures for proteins we had not crystallized, ProtiCelli gives us predicted images for proteins we have not stained. The fact that drug-induced changes emerge from morphology alone suggests the model has learned something genuinely deep about the relationship between cell shape and protein organisation.
\ud83d\udd2c T cell receptor (TCR) recognition of peptide-HLA complexes drives adaptive immunity, from fighting infections to rejecting tumors to triggering autoimmunity. But predicting which TCR will bind which peptide-HLA is notoriously difficult: the interaction interface is flexible, the sequence space is vast, and training data is sparse.
\n
StriMap is a unified deep learning framework from the Xavier lab at the Broad Institute and MIT that integrates physicochemical features, sequence context, and structural information at the recognition interface to predict TCR-peptide-HLA binding.
\n
\ud83e\uddec Rather than relying on sequence alone, StriMap models the 3D contact geometry between TCR CDR loops and the peptide-HLA surface. By combining structural priors with learned sequence representations, it captures binding patterns that pure sequence models miss, achieving state-of-the-art performance with improved generalizability across diverse datasets.
\n
\u26a1 The real validation came from application: the team screened 13 million peptides from 43,241 bacterial proteins to find molecular mimics relevant to ankylosing spondylitis (AS). Top candidates were experimentally validated to activate T cells expressing an AS-associated TCR, and one peptide showed enrichment in inflammatory bowel disease patients, suggesting shared microbial triggers between AS and IBD.
Structure-informed features at the binding interface improve generalization beyond the peptides and HLAs seen during training, breaking a critical bottleneck for TCR prediction models.
The AS screening demonstrates a complete workflow: computational prediction of 13M peptides narrowed to candidates that were experimentally validated to activate disease-relevant T cells.
\n
3\ufe0f\u20e3 Cross-Disease Molecular Mimicry
\n
A top validated microbial peptide showed enrichment in IBD patients, supporting the hypothesis that shared bacterial triggers drive immune cross-reactivity between AS and IBD.
\n
4\ufe0f\u20e3 Dual Application in Cancer and Autoimmunity
\n
The framework applies to both cancer neoantigen prediction and autoimmune antigen discovery, bridging two major areas of immunotherapy.
\n
\ud83d\udca1 Why This Is Cool
\n
TCR-antigen prediction has been one of the stubbornest problems in computational immunology. Most models fail to generalize beyond their training peptides. By bringing structural information into the prediction, StriMap breaks out of that limitation, and the AS/IBD discovery shows this is not just a benchmark improvement but a tool that finds real biology.
The best competitions, hackathons, and community challenges in AI x life sciences, curated weekly. Know something worth featuring? Reply and let us know.
A free, two-session workshop co-hosted by Manuel Corpas and Segun Fatumo, running entirely in Google Colab. No setup, no cost, just a browser. Session one covers variant interpretation using Ensembl VEP, ClinVar, and ACMG criteria. Session two runs a full GWAS pipeline with ClawBio, including polygenic risk scores and locus fine-mapping. Built to remove barriers for researchers anywhere in the world.
ELIXIR\u2019s annual international bioinformatics hackathon, running since 2018. 160+ participants, five days of collaborative coding on open bioinformatics infrastructure and tools. The call for project proposals opens March 16 and closes April 15 - so if you want to lead a project, that\u2019s your window.
\n
\n
Thanks for reading!
\n
\ud83d\udcac Get involved
\n
We\u2019re always looking to grow our community. If you\u2019d like to get involved, contribute ideas or share something you\u2019re building, fill out this form or reach out to me directly.
\n
Connect With Us
\n
Have questions or suggestions? We'd love to hear from you!
Thanks for reading Kiin Bio! Subscribe for free to receive new posts and support my work.
\n
\n\n\n
\n
\n","enclosure":{"link":"https://substackcdn.com/image/fetch/$s_!A5gQ!,f_auto,q_auto:good,fl_progressive:steep/https://substack-post-media.s3.amazonaws.com/public/images/d1b452f2-e701-44dd-ad83-454eebec3e43_1310x846.png","type":"image/jpeg"},"categories":[]},{"title":"\ud83e\uddecA Primer on AI Protein Design","pubDate":"2026-04-07 13:03:16","link":"https://newsletter.kiin.bio/p/a-primer-on-ai-protein-design","guid":"https://newsletter.kiin.bio/p/a-primer-on-ai-protein-design","author":"Natasha Kilroy","thumbnail":"","description":"The field went from predicting what proteins look like to designing ones that have never existed. Here's an intro to how.","content":"\n
Welcome back to Kiin Bio Weekly.
\n
For decades, designing a new protein meant years of directed evolution, rational engineering, and a lot of luck. You started from something nature already made and slowly nudged it toward what you wanted. The success rate was low. The timelines were long.
\n
That\u2019s changing fast. A wave of AI models, led by tools like RFDiffusion, ProteinMPNN, and AlphaFold, has opened up a fundamentally different approach: designing proteins from scratch, computationally, and getting functional molecules on the first try. In the last two years, AI-designed proteins have matched or outperformed naturally evolved ones in binding affinity, stability, and specificity, sometimes by significant margins.
\n
This primer covers what\u2019s actually happening, how the key methods work, and where the field is headed.
\n
\n
Keeping up with AI x life science news can get exhausting.
\n
It\u2019s scattered across LinkedIn, X, Substack, arXiv, Slack, newsletters... and you still somehow miss the things that actually matter. Too much noise, not enough signal.
\n
We\u2019re building something to fix that: a smarter, more powerful way to stay on top of what\u2019s actually relevant to you.
\n
But we want to build it with you, not just for you. Take 2 minutes to tell us what\u2019s missing. What you share will directly shape what we build, and you\u2019ll be the first to benefit from it.
The story starts with structure prediction. AlphaFold, released by DeepMind in 2020, solved a 50-year-old problem: predicting a protein\u2019s 3D structure from its amino acid sequence. That was transformative for understanding biology, but it didn\u2019t directly design new proteins. It told you what a sequence would fold into, not what sequence would give you the fold you wanted.
\n
The leap to design required inverting the problem. Instead of sequence \u2192 structure, the question became: what sequence would fold into a structure that does what I need?
\n
That inversion is what the current generation of tools enables.
\n
\n
\ud83e\udde9 The Key Methods
\n
There are three categories of AI tools driving protein design today, and they work best in combination.
\n
Structure generation creates new protein backbone shapes, the 3D scaffold. The breakthrough here is RFDiffusion, developed by David Baker\u2019s lab at the University of Washington. It uses diffusion models (the same class of generative AI behind image tools like DALL-E) applied to 3D coordinates. You specify what you want, a protein that binds a particular target, wraps around a small molecule, or presents a specific functional site, and the model generates backbone structures that satisfy those constraints. It\u2019s designing architectures that evolution never explored.
\n
Sequence design fills in the amino acid sequence for a given backbone. ProteinMPNN, also from Baker\u2019s lab, takes a 3D structure and predicts which amino acid sequences will fold into it stably. This is the bridge between a computational shape and something you can actually synthesise and test. It recovers native-like sequences with high accuracy and, critically, produces sequences that fold and function when tested experimentally.
\n
Structure prediction closes the loop. AlphaFold (and its open-source successor ESMFold from Meta) validates the designs by predicting whether the designed sequence will actually fold into the intended structure. If the predicted fold matches the designed backbone, confidence is high. If it doesn\u2019t, you iterate.
\n
The typical workflow today: RFDiffusion generates a backbone \u2192 ProteinMPNN designs sequences for it \u2192 AlphaFold confirms the fold \u2192 the best candidates go to the lab.
RFDiffusion can generate proteins for a range of design tasks, including binders for specific targets, symmetric assemblies, and scaffolds around functional motifs. From Watson et al., Nature (2023).
\n
The AI protein design workflow: a diffusion model generates novel backbone structures, a sequence design model fills in amino acids, and a structure prediction model validates the fold before experimental testing.
\n
\n
\u2699\ufe0f What\u2019s Actually Working
\n
The results from the last 18 months have been striking.
\n
De novo binders, proteins designed from scratch to bind a specific target, are now routinely achieving nanomolar affinity on the first experimental round, without any optimisation. A 2024 study from Baker\u2019s lab designed binders against a panel of therapeutic targets, including influenza and SARS-CoV-2, with success rates that would have been unthinkable five years ago.
\n
Protein design competitions are providing independent validation. Adaptyv Bio, a cloud lab for protein designers based in Lausanne, ran an open EGFR binder competition in 2024 that benchmarked AI design methods head-to-head with standardised experimental testing. The results showed a 5x improvement in design success rates compared to earlier approaches, with some AI-designed binders outperforming clinical antibodies.
\n
Stability is also improving. AI-designed proteins are increasingly more thermostable than their natural counterparts. They can be engineered to withstand higher temperatures and harsher conditions, which matters enormously for manufacturing and storage.
De novo protein binders designed by RFDiffusion. (b) Examples of computationally designed proteins (coloured) bound to their target proteins (blue), with arrows showing the design process. (c) A designed binder (pink) targeting Mdm2 (teal), a key cancer-related protein. (d) Experimental binding data confirming sub-nanomolar affinity - these proteins were designed from scratch and worked on the first try. From Watson et al., Nature (2023).
\n
\n
\ud83e\uddea The Validation Bottleneck
\n
Here\u2019s the catch: designing a protein computationally takes hours. Testing it experimentally still takes weeks to months.
\n
The field can now generate thousands of candidate designs per day. But each one needs to be synthesised, expressed, purified, and assayed to know if it actually works. That wet-lab step is the bottleneck, and it\u2019s where a lot of promising computational designs die, not because the design was wrong, but because the testing pipeline can\u2019t keep up.
\n
This is driving a new category of infrastructure: automated, high-throughput protein testing platforms that can validate designs at the speed AI generates them. The goal is a closed loop, design, test, learn, redesign, running continuously with minimal manual intervention. We\u2019ll be exploring this challenge in an upcoming deep dive with Adaptyv Bio, a cloud lab purpose-built for AI protein design validation. Stay tuned.
\n
Until that loop is fully closed, the practical throughput of AI protein design is limited not by the models but by the experiments.
\n
\n
\ud83d\udcca Beyond Binders
\n
Binding is the easiest thing to design for, because the objective is clear: does this protein stick to that target? But the field is pushing into harder problems.
\n
Enzyme design, creating proteins that catalyse specific chemical reactions, is significantly more challenging because function depends on precise atomic arrangements in the active site, not just overall shape. Early results are promising but success rates are lower than for binders.
\n
Multi-state design aims to create proteins that switch between conformations, molecular machines that respond to signals. This requires the model to optimise for multiple structures simultaneously, a much harder optimisation problem.
\n
Symmetric assemblies, protein cages, rings, and lattices, are being designed for drug delivery, vaccine design, and materials science. RFDiffusion has demonstrated the ability to generate novel symmetric architectures that self-assemble when tested experimentally.
\n
\n
\ud83d\udd2e Where This Is Going
\n
Three trends to watch.
\n
Generative models are getting multimodal. The next generation of design tools will jointly generate structure and sequence, rather than treating them as separate steps. Models that can reason about structure, sequence, dynamics, and function simultaneously will produce better designs faster.
\n
The data flywheel is spinning up. Every experimentally tested design, whether it works or not, generates training data that makes the next round of models better. Open repositories for protein design data are accelerating this. The more designs get tested, the faster the models improve.
\n
The design-test loop is tightening. As automated testing platforms scale, the gap between computational design and experimental validation will shrink. The long-term vision is protein design on demand: specify the function you want, get a validated molecule back in days rather than months.
\n
We\u2019re still early. Most AI-designed proteins are relatively simple, single-domain binders tested in controlled settings. The gap between designing a protein that binds a target in a tube and one that works as a drug in a patient remains enormous. But the trajectory is clear: the tools are getting better, faster, and more accessible. And the proteins they\u2019re producing are starting to work.
\n
\n
\ud83d\udcac Want to be featured in Kiin Bio Weekly?
\n
Each issue we speak directly with researchers, scientists, and builders working at the frontier of AI in life sciences. If you're working on something in this space and think it would resonate with our community, I'd love to hear from you - fill out this form or reach out to me directly.
\n
\n
Found this useful? Forward it to a colleague in drug discovery or protein design - it's the best way to help the newsletter grow.
Subscribe now to stay at the forefront of AI in Life Science. Every week: primers, deep dives, and direct conversations with the people building the field.
Welcome back to your weekly dose of AI news for Life Science!
\n
\n
Keeping up with AI x life science news can get exhausting.
\n
It\u2019s scattered across LinkedIn, X, Substack, arXiv, Slack, newsletters... and you still somehow miss the things that actually matter. Too much noise, not enough signal.
\n
We\u2019re building something to fix that: a smarter, more powerful way to stay on top of what\u2019s actually relevant to you.
\n
But we want to build it with you, not just for you. Take 2 minutes to tell us what\u2019s missing. What you share will directly shape what we build, and you\u2019ll be the first to benefit from it.
\ud83e\uddec RNA G-quadruplex-binding proteins regulate mRNA processing, localisation, and stress responses - but experimental detection alone can\u2019t scale. G4REP maps them across the entire human proteome.
\n
\ud83d\udd2c RNA G-quadruplexes act as post-transcriptional regulatory hubs, but identifying which proteins bind them is experimentally challenging. Classical approaches rely on known binding domains like zinc fingers, missing a wide range of RG4-interacting proteins.
\n
Researchers at the Universit\u00e0 degli Studi di Modena e Reggio Emilia and Sapienza Universit\u00e0 di Roma introduce G4REP, combining ESM-2 embeddings with LSTM neural networks to predict RG4-binding proteins at proteome scale.
\n
\ud83e\uddec G4REP analyses protein sequences for RG4 binding features: arginine-glycine-rich motifs, intrinsically disordered regions, and aromatic residues, identifying the short flexible motifs through which interactions typically occur.
\n
\u26a1 ~85% accuracy and AUROC of 0.91. Over 2000 candidate RG4-binding proteins identified across the human proteome, including 552 high-confidence hits.
RG4-binding proteins are enriched in stress granules linked to cancer and neurodegeneration. G4REP identifies candidates with ~85% accuracy, enabling reliable prioritisation of targets influencing RNA stability and translation.
\n
2\ufe0f\u20e3 Expanding RNA-Binding Protein Networks
\n
G4REP identified 2000+ candidate proteins including poorly characterised FAM families, expanding RNA regulation well beyond classical binding domains.
\n
3\ufe0f\u20e3 Predicting Functional Binding Regions
\n
G4REP pinpoints binding sites within proteins using a disorder-weighted residue score, highlighting short RGG-rich motifs as primary interaction sites.
552 high-confidence RG4-binding proteins localise to stress granules, supporting a coordinating role for RG4 interactions in RNA metabolism during cellular stress.
\n
\ud83d\udca1 Why This Is Cool
\n
G4REP opens up RG4 biology at a scale experimental methods alone cannot reach. It expands our map of RNA-protein interactions and surfaces previously hidden regulators and therapeutic candidates across the human proteome.
\ud83e\uddec Protein design has dozens of ML models. None of them talk to each other. Every real-world project still requires custom glue code, reformatting, and one-off pipelines. evedesign changes that.
\n
\ud83d\udd2c The design problems that matter most in protein engineering - conditional design under real-world constraints, multi-objective optimisation, and iterative lab-in-the-loop workflows - demand flexible, composable infrastructure that no single tool provides. Current ML methods are rarely interoperable and remain inaccessible to non-experts.
\n
Researchers at Harvard Medical School, Wellcome Sanger Institute, University of Cambridge, Seoul National University, Broad Institute, and collaborators built evedesign, a unified open-source framework that formalises conditional biosequence design in a method-agnostic way.
\n
\ud83e\uddec evedesign defines three composable operations - Generate, Score, and Transform - that work across any model type (MSA-based, LLM, inverse folding, de novo 3D). A standardised multi-level instance representation (sequence, embedding, and 3D structure simultaneously) lets outputs from one model feed directly into another without reformatting.
\n
\u26a1 Supports multi-objective optimisation, supervised and unsupervised model integration, and lab-in-the-loop iteration from the ground up. Interactive web interface at evedesign.bio takes users from target protein to orderable DNA. Demonstrated in antibody engineering, enzyme design, and natural enzyme discovery. MIT-licensed, FAIR-compliant, and self-hostable.
CDR design subject to multiple constraints (thermostability, deimmunisation, binding) using composable multi-objective pipelines - no custom code required.
\n
2\ufe0f\u20e3 Enzyme Design
\n
Combine generative models with scoring functions from different method families in a single workflow, iterating with experimental data as it comes in.
\n
3\ufe0f\u20e3 Lab-in-the-Loop Workflows
\n
Declarative, serialisable pipelines that can halt, incorporate new experimental results, and resume - built for iterative design rounds rather than one-shot prediction.
\n
4\ufe0f\u20e3 Accessibility for Non-Computational
\n
Researchers End-to-end web interface makes ML-driven protein design accessible to experimentalists without requiring programming or model-specific expertise.
\n
\ud83d\udca1 Why This Is Cool
\n
The bottleneck in protein engineering isn\u2019t any individual model - it\u2019s connecting them. evedesign is the first open-source framework that treats this as a first-class problem: standardised interfaces, composable workflows, and lab-in-the-loop iteration built in from the start. That\u2019s infrastructure the field has needed for years.
\ud83d\udd2c Peptide therapeutics are one of the fastest-growing drug modalities - they target flat, extended protein surfaces that small molecules can\u2019t reach. But computationally predicting how a peptide binds its target remains hard: peptides are flexible, often disordered, and fold upon binding. Classical docking methods struggle with all three.
\n
Researchers at Purdue University and Korea University College of Medicine review the full landscape of peptide-protein docking, from traditional physics-based approaches to the latest deep learning methods, covering 25+ tools across three generations.
\n
\ud83e\uddec The review organises methods into three categories: binding site predictors that guide or filter docking models, AlphaFold-based protocols for peptide-protein co-folding and refinement, and deep generative models (diffusion-based) that sample peptide conformations conditioned on a target structure.
\n
\u26a1 Diffusion models like RAPiDock and DiffPepDock are emerging as the most promising direction, handling peptide flexibility natively. AlphaFold-based methods work well for structured peptides but struggle with disordered ones. Major remaining gaps: limited training data, weak performance on long peptides, and almost no methods handle chemically modified peptides.
Maps the full computational toolkit for predicting peptide-protein binding modes - essential for designing peptide therapeutics against previously undruggable protein surfaces.
\n
2\ufe0f\u20e3 Masking Peptide Engineering
\n
Covers tools applicable to masking peptides that block drug binding sites until protease activation at the tumour - a growing immunotherapy strategy.
\n
3\ufe0f\u20e3 Practical Method Selection
\n
Table 1 catalogues 25+ methods with years, descriptions, and code availability links - a ready-made decision guide for choosing the right docking approach.
\n
4\ufe0f\u20e3 Open Challenges
\n
Identifies the three biggest gaps: training data scarcity, long/disordered peptide performance, and chemical modification handling - a roadmap for future method development.
\n
\ud83d\udca1 Why This Is Cool
\n
This is the review the peptide docking field needed. It doesn\u2019t just catalogue methods - it explains when each approach works, when it breaks, and what\u2019s missing. If you\u2019re working on peptide therapeutics, this is your decision guide for computational docking in 2026.
The best competitions, hackathons, and community challenges in AI x life sciences, curated weekly. Know something worth featuring? Reply and let us know.
\n
\ud83c\udfd7\ufe0f Hackathon Highlight
\n
The Elnora x Monomer Bio AI scientist hackathon brought together 43 builders for 24 hours of agent-driven lab automation. The winning team connected Elnora\u2019s protocol generation agent to a robotic arm that autonomously imaged cells, assessed confluency at 70%, generated a splitting protocol, executed it, and re-imaged to confirm results - a fully closed AI-to-robot lab loop built from scratch overnight. 6 out of 7 teams incorporated Elnora into their workflows.
Last weekend, Edinburgh hosted its first ever biohackathon, and the numbers speak for themselves.
\ud83c\udfaf 350 applications. 110 selected participants. Over 95% attendance rate, which for a free hackathon is exceptional (typical rates sit around 40-50%).
Seven challenges spanned three tracks: academic, industry and non-coder. The two standouts were the bio-business and Pacifico Biolabs (Genome-Scale Metabolic Modelling Tool) track, which saw the highest participation with 8 and 7 teams respectively.
\ud83d\udd2c One project that caught everyone's attention: FilamentTracker. A team built a fully functional web tool with its own custom domain that detects and tracks protein filaments in microscopy images, showing filament length and area over time. In 48 hours. Judges flagged it as potentially publishable if stress-tested across other organisms.
What made this event different was the deliberate push for interdisciplinary collaboration. Teams were encouraged to mix engineers, biologists, data scientists and non-coders. Not everyone landed in a mixed team, but the feedback was clear: those who did got the most out of it.
All submissions are open source on DevPost (https://lnkd.in/gjpQwXXR). Next up from the organisers: peer-led company creation workshops to help winning teams take their projects further.
\ud83d\ude80 Scotland's first biohackathon, built from a blank Google Drive folder by an unpaid volunteer team. If this is what version one looks like, version two is going to be something special!
ELIXIR\u2019s annual international bioinformatics hackathon, running since 2018. 160+ participants, five days of collaborative coding on open bioinformatics infrastructure and tools. The call for project proposals opens March 16 and closes April 15 - so if you want to lead a project, that\u2019s your window.
\n
\n
Thanks for reading!
\n
\ud83d\udcac Get involved
\n
We\u2019re always looking to grow our community. If you\u2019d like to get involved, contribute ideas or share something you\u2019re building, fill out this form or reach out to me directly.
\n
Connect With Us
\n
Have questions or suggestions? We'd love to hear from you!
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!ol8B!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c717a0a-e0a9-4a33-be93-9219ccdd0fff_1464x360.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!ol8B!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c717a0a-e0a9-4a33-be93-9219ccdd0fff_1464x360.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!YBZa!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fca5c5956-45e3-4575-beea-b479eba28539_2048x1179.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!6YA-!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fccd7e09e-6e05-4bcb-a064-e63d3650692e_2048x1471.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!RGzT!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd1483b5c-e026-4cda-8f60-af71536f0fa1_683x1024.jpeg\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!Am9s!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9ddf4e32-3d87-41a6-8973-e0f91b110778_2048x1219.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!CBlc!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc4ef6aee-a6a5-400a-8acf-3ac39570bd33_2048x1366.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!vCIn!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F55811f4c-5198-43c2-b45f-cdbf193581ed_2817x1512.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!sUtc!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4de33a82-6a53-4c40-ae5a-0b7555b56249_2978x1490.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!paEG!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F50ce573e-690e-42f3-8fe6-f6940df02b4b_1352x1454.png\" "\\"\\"")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!P7mX!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa4a32f5d-c8d4-4cbb-a037-f51fa018eaf2_677x913.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!BNn9!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F27e4bf4a-435a-42e6-b656-c39d56dd9443_1198x350.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!jBOE!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F62a7ce88-89f4-4c44-a1f1-e2afaf75eaa2_1276x974.png\")
![\"diagram\"](\"https://substackcdn.com/image/fetch/%24s_!PI1b!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff5793849-51c2-4b5b-9fd2-714e5c1e046e_800x535.jpeg\" "\\"diagram\\"")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!27Vz!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F14ea94be-5c37-4df4-b2f3-f1ab9af900ae_862x1212.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!QOmd!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fced8f805-1992-4125-bdcc-1e81e269f7cf_1188x394.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!3Xhc!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb153eb21-6289-4754-87b8-122b06ee7d45_1724x1482.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!sQWC!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2c9ba435-b0b1-4689-9c5a-b0fff42432f6_1476x1310.png\")
![\"diagram\"](\"https://substackcdn.com/image/fetch/%24s_!QRPj!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F751000d4-bf60-4197-9199-93b6261f88bb_1166x416.jpeg\" "\\"diagram\\"")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!A2u0!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd2332d09-2d3e-41be-b750-7d5d84b57f73_1266x780.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!GY52!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F07292c19-f37c-4c14-91b6-baa7b3b5f943_1462x548.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!TSot!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1ede8148-289d-4712-aecc-3206b40efb90_1920x574.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!bKTB!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fde678518-64e3-4576-b8be-a259a091688b_572x592.png\")
![\"No](\"https://substackcdn.com/image/fetch/%24s_!s8aW!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc6f409b2-7d41-44ad-a625-8d1ad211ec07_796x236.jpeg\" "\\"No")
![\"diagram,](\"https://substackcdn.com/image/fetch/%24s_!J9Ni!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1d7feb9d-f756-4512-8560-49dcd352418d_800x710.jpeg\" "\\"diagram,")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!y9Y3!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffe3ed73e-6e87-4491-b29c-1b521e230af2_1084x1168.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!6ZQB!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffbe4d499-de75-40fd-8f31-5a1199418a77_1021x509.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!a861!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7294589d-d400-464d-a433-656f8abd11a9_1024x511.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!gYeb!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fce905d89-e24c-45a5-924a-2da65ef6ff92_2048x1365.jpeg\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!A5gQ!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd1b452f2-e701-44dd-ad83-454eebec3e43_1310x846.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!sbHR!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F0773db4c-5ca3-4fd7-937a-61afd5497c3f_1230x1044.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!J-L3!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F42812c9c-f2e2-440a-8c50-38754726b1a9_1592x1080.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!l4Mp!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F35c6b79f-8a0d-4d6b-9da2-39602a2122c3_672x490.png\")
![\"graphical](\"https://substackcdn.com/image/fetch/%24s_!ylYR!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4ed2c930-99fb-4575-8027-44da78fb2874_800x433.jpeg\" "\\"graphical")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!KpsN!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc83ce0bc-683b-4a92-823d-300250703ede_1156x1014.png\")
![\"\"](\"https://substackcdn.com/image/fetch/%24s_!q2mg!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F7c79b3ef-2f94-427d-ba67-dab5c362d5ef_1574x838.png\")