Representation Learning Duck
I have found a principled way to recover known and find new drug targets using the t-cell perturbation data for any disease whilst highlighting the interpretability of how those drug targets affect the proceeding biologi
Core trick
Learns interpretable gene programs from CD4 T-cell Perturb-seq, links programs causally, and runs the full discovery pipeline through an agent.
Description
I have found a principled way to recover known and find new drug targets using the t-cell perturbation data for any disease whilst highlighting the interpretability of how those drug targets affect the proceeding biological programs.
I have borrowed methods from the field of causal representation learning to build relationships between the latent programs and found how perturbations would cascade down programs and how it effects each of them in a connected manner.
This helped with the interpretability of the results of novel drug targets found for SLE and other potential diseases.
Moreover, I have also built an agent that can be used to run the whole pipeline with any disease in mind, end to end, providing the results and potential novel drug targets.
d-setiawan/Interpretable-TCell-Representation-Learning★ 00 forksHTMLCreated Jul 13, 2026 · last push Jul 13, 2026
README excerpt
# Interpretable T-cell Representation Learning
**An interpretable gene-program model of CD4⁺ T-cell perturbation for drug-target discovery —
driven by a Claude agent.**
From a genome-scale CD4⁺ T-cell CRISPRi Perturb-seq screen, we learn **32 reproducible, interpretable
gene programs** — the *most* interpretable of six representation-learning methods — and use them to
recover known regulators and approved drug targets, nominate novel candidates by **functional–genetic
convergence**, and drive **AutoTarget**, an autonomous Claude agent that produces self-critiqued target
dossiers. Every claim is benchmarked against explicit baselines, with an honest account of what is and
isn't causally identifiable.
<p align="center">
<a href="https://d-setiawan.github.io/Interpretable-TCell-Representation-Learning/">
<img src="docs/preview.png" width="840" alt="The Latent Causal Atlas — interactive dashboard">
</a>
<br>
<a href="https://d-setiawan.github.io/Interpretable-TCell-Representation-Learning/"><b>▶ Launch the interactive dashboard →</b></a>
· trace any knockdown → gene programs → drug targets, live in your browser
</p>
### Quick links
- 🔬 **[Live interactive dashboard](https://d-setiawan.github.io/Interpretable-TCell-Representation-Learning/)** — the fastest way to get a feel for the project: trace any knockdown through the gene programs, see how it maps to drug targets, and browse the AutoTarget agent dossiers. *(One self-contained page