AutoScreen: AI Co-Scientist System for Target Discovery in Functional Genomics
- 1Department of Pathology, Department of Genetics, Cancer Biology Program, Stanford University, Stanford, CA 94305, USA
Abstract
Target discovery in functional genomics remains largely manual and time-consuming, lacking systematic tools for efficient and reproducible gene-level hypothesis generation. We introduce AutoScreen, an AI co-scientist system supporting target discovery through both Pre-screen Design, which constructs perturbation libraries de novo from free-text research descriptions, and Post-screen Analysis, which re-ranks experimental screen hits by integrating statistical scores with biological context. AutoScreen leverages the complementary strengths of multiple specialist agents that perform deep research into multi-modal evidence, information restructuring, parallel searches across >26 biomedical databases, evidence synthesis, and target review to provide transparent, explainable gene prioritization with pipeline provenance. Across 320 genome-scale CRISPR screens as expert-curated benchmarks, AutoScreen achieved a ~19% increase in validated-hit recovery among its top 100 predictions, relative to the strongest agent baseline, and required a 1.2-fold smaller library to recover the same number of hits at the top-500 reference point. AutoScreen reached mean average precision more than two orders-of-magnitude above random baseline. Further, we validated AutoScreen in cancer immune-evasion case studies focusing on natural killer (NK) and T-cell therapeutics. AutoScreen recovered NK-resistance genes that were initially lower-ranked in a leukemia screen, moving MUC1, PDPN, and LRRC15 from raw ranks of 118, 81, 1384 to 5, 44, 659, respectively. In follow-up tumor killing assay with primary human NK cells, individual perturbation validated all three hits successfully. Next, in prospective benchmarking across two cytotoxic T-cell-killing screens, AutoScreen recovered 77.1% of ground-truth hits called by the consensus of two gold-standard analysis pipelines (FDR<0.1), achieving 8% improvement over the strongest general-purpose LLM agent baseline. Finally, we re-analyzed >500 public datasets to construct the AutoScreen Resource Hub, a growing knowledge base of pre-computed, annotated reports for CRISPR screens, RNA-seq differential expression, and gene-level UK Biobank genome-wide association studies. This agentic AI approach enables real-time genomics benchmark construction that are continuously updated, expanded, for evaluating frontier AI co-scientists. Overall, AutoScreen enables AI-powered target discovery to be more auditable, scalable, extensible, and reproducible.
Related
- A Closed-Loop Robot Scientist for Autonomous Biological Discovery
- Protein co-scientist aided simulations of the structural and functional impact of O-GlcNAc Transferase TPR domain deletions and amino acid substitutions associated with X-Linked Intellectual Disability
- Comparative Single-Cell Profiling of CRISPR Knockout and Interference Defines Modality-Specific Strengths in Functional Genomics
- Refining the genetic architecture of myocardial infarction through integrative post-genome-wide association study functional genomics analysis
- Comparative genomics reveals a genomic and functional continuum among secondary replicons in Halobacteriota
BibTeX
@misc{qu2026autoscreen,
title = {{AutoScreen: AI Co-Scientist System for Target Discovery in Functional Genomics}},
author = {Y. Qu and X. Liu and X. Wang and M. Chen and X. Luo and L. Lyu and M. Yin and J. Hui and D. Yin and R. Dinesh and L. Qiu and K. Huang and H. Wang and S. Tong and H. Cousins and R. Feng and O. Martinez and J. Zhang and T. Chen and R. Altman and J. Leskovec and A. Regev and M. Wang and L. Cong},
howpublished = {bioRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.06.749678},
url = {https://www.biorxiv.org/content/10.64898/2026.09.06.749678v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - AutoScreen: AI Co-Scientist System for Target Discovery in Functional Genomics
AU - Y. Qu
AU - X. Liu
AU - X. Wang
AU - M. Chen
AU - X. Luo
AU - L. Lyu
AU - M. Yin
AU - J. Hui
AU - D. Yin
AU - R. Dinesh
AU - L. Qiu
AU - K. Huang
AU - H. Wang
AU - S. Tong
AU - H. Cousins
AU - R. Feng
AU - O. Martinez
AU - J. Zhang
AU - T. Chen
AU - R. Altman
AU - J. Leskovec
AU - A. Regev
AU - M. Wang
AU - L. Cong
PY - 2026
DA - 2026/09/06
PB - bioRxiv
DO - 10.64898/2026.09.06.749678
UR - https://www.biorxiv.org/content/10.64898/2026.09.06.749678v1
AB - Target discovery in functional genomics remains largely manual and time-consuming, lacking systematic tools for efficient and reproducible gene-level hypothesis generation. We introduce AutoScreen, an AI co-scientist system supporting target discovery through both Pre-screen Design, which constructs perturbation libraries de novo from free-text research descriptions, and Post-screen Analysis, which re-ranks experimental screen hits by integrating statistical scores with biological context. AutoScreen leverages the complementary strengths of multiple specialist agents that perform deep research into multi-modal evidence, information restructuring, parallel searches across >26 biomedical databases, evidence synthesis, and target review to provide transparent, explainable gene prioritization with pipeline provenance. Across 320 genome-scale CRISPR screens as expert-curated benchmarks, AutoScreen achieved a ~19% increase in validated-hit recovery among its top 100 predictions, relative to the strongest agent baseline, and required a 1.2-fold smaller library to recover the same number of hits at the top-500 reference point. AutoScreen reached mean average precision more than two orders-of-magnitude above random baseline. Further, we validated AutoScreen in cancer immune-evasion case studies focusing on natural killer (NK) and T-cell therapeutics. AutoScreen recovered NK-resistance genes that were initially lower-ranked in a leukemia screen, moving MUC1, PDPN, and LRRC15 from raw ranks of 118, 81, 1384 to 5, 44, 659, respectively. In follow-up tumor killing assay with primary human NK cells, individual perturbation validated all three hits successfully. Next, in prospective benchmarking across two cytotoxic T-cell-killing screens, AutoScreen recovered 77.1% of ground-truth hits called by the consensus of two gold-standard analysis pipelines (FDR<0.1), achieving 8% improvement over the strongest general-purpose LLM agent baseline. Finally, we re-analyzed >500 public datasets to construct the AutoScreen Resource Hub, a growing knowledge base of pre-computed, annotated reports for CRISPR screens, RNA-seq differential expression, and gene-level UK Biobank genome-wide association studies. This agentic AI approach enables real-time genomics benchmark construction that are continuously updated, expanded, for evaluating frontier AI co-scientists. Overall, AutoScreen enables AI-powered target discovery to be more auditable, scalable, extensible, and reproducible.
ER -