The hottest Computational Biology Substack posts right now

And their main takeaways
Category
Top Technology Topics

How to Design Antibodies

Asimov Press • 754 implied HN points • 11 Mar 26
🔬 Science Computational Biology
  1. AI models now let researchers design antibody binders on the computer, greatly reducing the experimental search effort needed to find promising candidates.
  2. There is a practical five-step pipeline — pick a target, prepare or predict its structure, run design tools, filter candidates, and validate in the lab — which uses public tools but typically costs thousands of dollars.
  3. Design success is highly target-dependent and improving affinity, specificity, and drug-like properties remains difficult and costly, but AI makes it realistic to engineer more complex, multi-property binders going forward.

Scent, In Silico

Asimov Press • 432 implied HN points • 16 Feb 26
🔬 Science Computational Biology
  1. Smell is an ancient, highly combinatorial sense driven by hundreds of receptor types, so odors come from complex mixtures and are inherently subjective.
  2. New computational tools like graph neural networks create odor embeddings that map molecules into a perceptual space, letting machines predict smells and design novel odorants.
  3. Digitizing scent promises faster fragrance discovery, diagnostics, safer repellents, and more sustainable synthetic alternatives, while also raising questions about authenticity and how we value natural versus machine-made ingredients.

Getting Back To It

Rough Diamonds • 67 implied HN points • 26 Feb 26
🕹 Technology Computational Biology
  1. A major life transition — having a baby and actively searching for AI-related roles — is prompting a return to team-based work and a desire to re-engage with public writing.
  2. Hands-on AI work is central: building personal tools like a life-tracker and a personal CRM, analyzing LLM usage, and experimenting with coding agents and AI-for-science applications.
  3. Nuanced, pragmatic views on AI and life: supportive of useful AI but sympathetic to critics, wary of AI-assisted creative work, expecting closed-loop lab automation to grow but not yet ubiquitous, and valuing simplicity, human-centered practices, and taste-driven giving.

Machine Learning-Assisted Directed Evolution with Bruce Wittmann

Lever • 19 implied HN points • 16 Oct 24
🔬 Science Computational Biology
  1. Bruce Wittmann's journey in science started from pre-med and led him to research at notable institutes like Caltech.
  2. He worked on machine learning to improve protein engineering, building tools that can help many people in the field.
  3. His collaboration with renowned scientists and contributions to published research highlight the exciting potential in protein design and computational biology.

Self-Replicating Programs: Life's Origin Story Told Through Brainfuck and Assembly

Gonzo ML • 441 implied HN points • 16 Dec 25
🔬 Science Computational Biology
  1. Self-replicating programs can spontaneously emerge from random code when programs interact and rewrite each other, without hand-built ancestors or an explicit fitness function.
  2. This emergence happens across many computational substrates and spatial setups (brainfuck variants, Forth, Z80, i8080, 0D/1D/2D, long tapes), though some languages resist, so language features and locality shape how and how fast replicators appear.
  3. The system shows a clear phase transition — complexity and copyable tokens spike as replicators take over — and the resulting dynamics (competition, coexistence, niche creation) mirror ecological and origin-of-life concepts.
Get a weekly roundup of the best Substack posts, by hacker news affinity:

Sneaky AI chess, cancer cell deception, & plant ballistic strategies

The Strategy Toolkit • 26 implied HN points • 26 Jan 26
🔬 Science Computational Biology
  1. AI systems can be tricked into accepting false rule changes and making illegal moves, highlighting real vulnerability to deception.
  2. Public AI competitions on social media turn technical failures into vivid, easy-to-follow lessons about strategic behavior.
  3. Watching AI-versus-AI interactions gives strategists practical insights into trust, adversarial tactics, and how to build more robust systems.

Automating Biology

A Biologist's Guide to Life • 29 implied HN points • 26 Jan 26
🔬 Science Computational Biology
  1. Living systems are layers of metabolic machines — from genes and proteins to cells, tissues, and organisms — that act like modular, self-replicating components we can study and engineer.
  2. Physical automation (robotic labs, cloud labs) and digital automation (AI-driven biodesign and structure prediction) can make experiments much cheaper, higher-throughput, and faster, enabling far more data and quicker innovation.
  3. Widespread automation is limited by trust, data security, and the need for flexibility as methods evolve, so modular, autonomous lab systems and careful governance are needed to realize its promise.

agent.bio

LatchBio • 63 implied HN points • 09 Dec 25
🔬 Science Computational Biology
  1. An interactive sandbox hosts natural-language agents tailored to five major spatial biology platforms (Takara Seeker, Vizgen MERFISH, AtlasXOmics DBiT-seq, 10X Xenium, and 10X Visium) so scientists can run end-to-end spatial analyses.
  2. Agents operate in two modes—"proactive" for automated runs and "step-by-step" for frequent check-ins—and users should prefer step-by-step for important work because the sandboxes were built for specific datasets and may not generalize perfectly.
  3. Video demos show these agents can ingest raw outputs, run QC, clustering, differential tests, and annotate spatial features across diverse biological problems, and the roadmap focuses on benchmarks, purpose-built infrastructure, and tech-specific heuristics to make agents reliable for scientific decisions.

SpatialBench: Can Agents Analyze Real-World Spatial Biology Data?

LatchBio • 41 implied HN points • 26 Dec 25
🔬 Science Computational Biology
  1. SpatialBench is a realistic suite of 146 verifiable spatial biology problems across five platforms and seven task types that recreates real analyst workspaces using snapshots of data and images.
  2. Current agent models perform poorly overall (roughly 20–38% accuracy) and vary widely by task and platform, and the choice of execution harness or wrapper can change outcomes as much as changing the base model.
  3. Inspecting agent trajectories reveals clear failure modes and productive strategies, showing that detailed traces help explain performance and that benchmarks like this are a practical first step toward engineering agents that can reliably automate spatial biology analysis.

Cancer-Selective, Pan-Essential Targets from DepMap

Rough Diamonds • 20 implied HN points • 09 Jan 26
🔬 Science Computational Biology
  1. Large-scale DepMap screening can flag genes whose knockout strongly harms many cancer cell lines but not immortalized "normal" lines, yet these results are limited by dataset definitions (many hits fall on DepMap's "pan-essential" list) and by the poor representation of healthy human tissues in culture, so experimental validation is needed.
  2. The top candidates include both familiar chemotherapy targets and new leads: some targets already have clinical-stage inhibitors or ADCs (e.g., TFRC, NMT1), while others (e.g., YRDC, SEPHS2, PHF5A, ADSL) are preclinical or underexplored and could be druggable by different modalities.
  3. LLM-generated code (Claude Code) made the project fast and reproducible, but agent-produced code can silently change behavior or omit checks, so careful human review, testing, and follow-up biological experiments are essential.

The Evolutionary Scales of Life

A Biologist's Guide to Life • 15 implied HN points • 27 Dec 25
🔬 Science Computational Biology
  1. Ecological patterns depend on the spatial, temporal, and evolutionary scale you examine; changing the scale can reveal or hide important patterns.
  2. Phylofactorization is an algorithm that finds edges or clades in a phylogenetic tree that best explain differences in traits or ecological patterns, letting you partition life at the scales that matter for a given question.
  3. There is no single correct species or taxonomic scale; instead choose or infer the lineage-level scales that match your question, and tree-based partitioning can also reveal relevant scales in non-biological hierarchical systems.

Horrors Of The Theoretical Maximum

ASeq Newsletter • 14 implied HN points • 30 Oct 24
🕹 Technology Computational Biology
  1. Vendors sometimes quote theoretical maximums for data output, which can be misleading. It's important to understand that these numbers might not reflect actual performance.
  2. Comparing different technologies can be complicated because they have different specifications and capabilities. Each technology, like PacBio, Oxford Nanopore, and Illumina, has its unique strengths and limitations.
  3. In the real world, the difference between what is theoretically possible and what is actually achieved can be significant. This means we should be cautious and not rely solely on theoretical figures.

GPU-enabled Bioinformatics

LatchBio • 9 implied HN points • 06 Nov 24
🕹 Technology Computational Biology
  1. Bioinformatics is moving towards using GPUs to speed up data processing. This change can save a lot of time and money for researchers.
  2. New molecular techniques generate massive amounts of data that take too long to analyze without faster systems. Using GPUs can make these processes much quicker, especially for large datasets.
  3. There are now cloud platforms that make it easier to use GPU technology without needing special expertise or expensive hardware. This helps more teams access advanced analysis tools.

TechBio Adventures

Kesav’s Lab • 1 HN point • 16 Feb 24
🕹 Technology Computational Biology
  1. TechBio combines biology and technology to make advancements in healthcare. This approach allows for faster and more efficient drug development.
  2. Understanding DNA and using software tools are key parts of TechBio. This lets us design new biological systems to solve complex problems.
  3. There are two main areas in TechBio: industrial and clinical applications. Both aim to improve health outcomes and automate biological processes.