Medea: An AI agent for therapeutic reasoning across biological contexts

Target and Antibody Discovery via AI

my Ph.D. Thesis: transformers on single cell RNA-seq as large cell models

A reproducible single-cell workflow for human liver fibrosis, MASH, and cirrhosis target discovery.

End-to-end GNN pipeline for drug–target interaction prediction and target prioritization

Insilico Medicine is a clinical-stage biotechnology company applying generative AI and robotic automation to drug discovery and development. Its Pharma.AI software suite includes PandaOmics (target discovery from multi-omics data), Chemistry42 (generative chemistry with retrosynthesis, property prediction, and ADMET), Generative Biologics…

Computational target discovery + antigen-specific therapeutic design for inflammatory/immune disorders (EGIDs, celiac, IBD, autoimmune). Bundles the omics-mining, pMHC-epitope, and systematic-review methodology with DOI-verification and presentation-

A reproducible multi-omics bioinformatics workflow for investigating extracellular matrix-associated molecular signatures in pancreatic ductal adenocarcinoma.

Predicts CRISPR gene dependencies across 1,066 DepMap cancer models and ranks selective therapeutic hypotheses. Precision@10 0.965 on held-out models, validated against an independent Sanger screen.

Parkinson's disease target, compound, omics, stem-cell validation and knowledge-graph benchmark dashboard

TargetLens / 靶点梳理助手：可追溯的肿瘤靶点研读与证据决策工作台

TargetIntel-IO is an explainable, rule-based tool that ranks and classifies candidate targets in anti-PD-1-resistant melanoma by therapeutic intent: antibody/IO-combination, biomarker, or small-molecule strategy, with evidence-based reasoning and benchmark-checked accuracy behind every call.

Disease-agnostic methodology stack for computational target discovery and antigen-specific therapeutic design in inflammatory/immune disorders (EGIDs, celiac, IBD, autoimmune).

AI-powered cancer target discovery agent for precision oncology research. Input genomic variants → find matched drugs via CIViC/DGIdb/PubMed → rank by evidence. Built with LangGraph + DeepSeek API + Streamlit.

End-to-end, human-in-the-loop therapeutic drug-development pipeline for Claude Science — archetype-adaptive across common and genetic/rare disease.

IL-17 target prioritization for Plaque Psoriasis - FAIR Nextflow workflow and dashboard integrating public HTS/qHTS assay data, psoriasis omics, proteomics, protein features, structural context, and grounded LLM evidence summaries.