6 min read

Foundation models eat single-cell biology

Foundation models eat single-cell biology
Nº 01 · The Lede X Field report

Tabula predicts gene regulation, aging

Tabula predicts gene regulation, aging
Fig. IX · Filed 21 Jul 2026.

Tabula, a single-cell foundation model, was unveiled by Xiaojie Qiu's group with a privacy-preserving tabular training scheme — labs contribute data without exposing raw counts. The model targets gene-regulatory inference and aging trajectories, two areas where scRNA-seq foundation models have historically underperformed task-specific tools. Trained across federated cohorts, Tabula reframes what a virtual-cell model can promise: predictive regulatory dynamics, not just embedding quality, with a data-governance story attached.

Read the source

OpenAI turns agents on safety
Fig. IIX · Filed 21 Jul 2026.
Nº 02 X Agents · Infrastructure

OpenAI turns agents on safety

OpenAI introduced GPT-Red, an agent flywheel where current models red-team the next generation's safety evaluations. The pitch: automated adversarial testing scales faster than human review, and the same agent scaffolding driving capability gains now drives safety gains. Moves adversarial evaluation from a human bottleneck to a compute-bound process — reshapes what regulators and clinical deployers can reasonably demand before biomedical agents ship.

Read more
Anthropic funds rare-disease AI
Fig. IIIHacker News · Filed 21 Jul 2026.
Nº 03 Hacker News Field report

Anthropic funds rare-disease AI

Anthropic opened a targeted AI for Science call for rare genetic disease research, offering up to $50,000 in Claude credits over six months per accepted team. Small dollars by grant standards, but the signal matters: rare disease becomes a named vertical where a frontier lab wants a research community — anchors a reference funder for LLM-driven rare-disease work that had been scattered across one-off pilots.

Read more
Also Filed · Five Briefs from the queue
Nº 04 bioRxiv Field report

ProtSyntax decodes PTM grammar

ProtSyntax, a protein language model trained to read post-translational modification syntax, predicts modification sites and their functional consequences jointly. PTMs have been the blind spot of protein LMs — sequence-only models miss most of the regulatory action. Closes a long-standing gap in what protein foundation models can meaningfully claim about function.

Read
Nº 05 arXiv Field report

GigaPath-Flash speeds up pathology

GigaPath-Flash and GigaTIME-Flash compress Microsoft's pathology foundation models for whole-slide and tumor-microenvironment analysis at a fraction of the compute. Moves gigapixel pathology inference from cluster-scale to workstation-scale — narrows the deployment gap between foundation-model pathology and the hardware most cancer centers actually have.

Read
Nº 06 arXiv Field report

LLMs stumble on 3D binding

A new benchmark asks whether language models can reason about molecular binding under spatial constraints. Answer: not really — LLMs handle 2D chemistry text well but fail when 3D geometry decides the answer. Anchors a hard ceiling on text-only chemistry claims — the same wall sequence models hit on 3D genome folding — and pushes the field back toward structure-aware architectures for drug design.

Read
Nº 07 bioRxiv Field report

CpGPT models DNA methylation

CpGPT extends the foundation-model template to CpG methylation, giving epigenetic clocks and methylation-based diagnostics a pretrained base to fine-tune from. Methylation joins protein, single-cell, and pathology as modalities with their own foundation model.

Read
Nº 08 Anthropic Agents · Infrastructure

Clay Seal proposes agent identity

Clay Seal Identity, an open-source project, pitches cryptographic accountability for agent actions — every tool call signed, every output traceable. Early and quiet on HN, but the direction matters: agent provenance is drifting toward a vendor criterion for anything touching clinical or regulated data.

Read

Reply with your discoveries. A human reads them. Forward freely.

Agentic Discovery  ·  Nº 60  ·  21 Jul 2026

Editor's Note

Foundation-model Tuesday: single-cell, pathology, methylation, and PTMs all get their own — the era of one-model-per-modality is arriving fast.

 

Nº 01 · The Lede  —  X  —  Field report

Tabula predicts gene regulation, aging

Tabula predicts gene regulation, aging

Fig. I  X · Filed 21 Jul 2026.

Tabula, a single-cell foundation model, was unveiled by Xiaojie Qiu's group with a privacy-preserving tabular training scheme — labs contribute data without exposing raw counts. The model targets gene-regulatory inference and aging trajectories, two areas where scRNA-seq foundation models have historically underperformed task-specific tools. Trained across federated cohorts, Tabula reframes what a virtual-cell model can promise: predictive regulatory dynamics, not just embedding quality, with a data-governance story attached.

Read the source →

Why it matters

Single-cell foundation models have spent two years failing to beat specialist baselines on regulation and perturbation — a credible predictive result with built-in privacy plumbing resets both the capability ceiling and the data-sharing template that virtual-cell consortia will now be measured against.

 

Nº 02  —  X  —  Agents · Infrastructure

OpenAI turns agents on safety

Fig. II  X · Filed 21 Jul 2026.

OpenAI turns agents on safety

OpenAI introduced GPT-Red, an agent flywheel where current models red-team the next generation's safety evaluations. The pitch: automated adversarial testing scales faster than human review, and the same agent scaffolding driving capability gains now drives safety gains. Moves adversarial evaluation from a human bottleneck to a compute-bound process — reshapes what regulators and clinical deployers can reasonably demand before biomedical agents ship.

Read more →

 

Nº 03  —  Hacker News  —  Field report

Anthropic funds rare-disease AI

Fig. III  Hacker News · Filed 21 Jul 2026.

Anthropic funds rare-disease AI

Anthropic opened a targeted AI for Science call for rare genetic disease research, offering up to $50,000 in Claude credits over six months per accepted team. Small dollars by grant standards, but the signal matters: rare disease becomes a named vertical where a frontier lab wants a research community — anchors a reference funder for LLM-driven rare-disease work that had been scattered across one-off pilots.

Read more →

 

Also Filed  ·  Five Briefs from the queue

Nº 04  —  bioRxiv  —  Field report

ProtSyntax decodes PTM grammar

ProtSyntax, a protein language model trained to read post-translational modification syntax, predicts modification sites and their functional consequences jointly. PTMs have been the blind spot of protein LMs — sequence-only models miss most of the regulatory action. Closes a long-standing gap in what protein foundation models can meaningfully claim about function.

Read →

Nº 05  —  arXiv  —  Field report

GigaPath-Flash speeds up pathology

GigaPath-Flash and GigaTIME-Flash compress Microsoft's pathology foundation models for whole-slide and tumor-microenvironment analysis at a fraction of the compute. Moves gigapixel pathology inference from cluster-scale to workstation-scale — narrows the deployment gap between foundation-model pathology and the hardware most cancer centers actually have.

Read →

Nº 06  —  arXiv  —  Field report

LLMs stumble on 3D binding

A new benchmark asks whether language models can reason about molecular binding under spatial constraints. Answer: not really — LLMs handle 2D chemistry text well but fail when 3D geometry decides the answer. Anchors a hard ceiling on text-only chemistry claims — the same wall sequence models hit on 3D genome folding — and pushes the field back toward structure-aware architectures for drug design.

Read →

Nº 07  —  bioRxiv  —  Field report

CpGPT models DNA methylation

CpGPT extends the foundation-model template to CpG methylation, giving epigenetic clocks and methylation-based diagnostics a pretrained base to fine-tune from. Methylation joins protein, single-cell, and pathology as modalities with their own foundation model.

Read →

Nº 08  —  Anthropic  —  Agents · Infrastructure

Clay Seal proposes agent identity

Clay Seal Identity, an open-source project, pitches cryptographic accountability for agent actions — every tool call signed, every output traceable. Early and quiet on HN, but the direction matters: agent provenance is drifting toward a vendor criterion for anything touching clinical or regulated data.

Read →

 

· · ·

Reply with your discoveries. A human reads them. Forward freely.