Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-1820
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-18T10:43:01.336258Z
status: published
lens: g_space
sector: science
headline: Omics data discovery agents: Agent-supported retrieval, reanalysis, and synthesis of published omics data
dek: The biomedical literature contains a vast collection of omics studies, yet most published data remain functionally inaccessible for computational reuse. When raw data are deposited in public repositories, essential information for reproducing reported results is dispersed across main text, supplementary files, and code repositories, and in the rarer cases where intermediate data (e.g., protein abundance files) are shared, their location is irregular. Here we present an agentic framework for the agent-supported r…
gain_title: LLM agents with containerized tools completed five proteomics and RNA-seq reanalyses that closely reproduced authors' deposited abundances, supporting automated reuse of published omics data.
problem_title: (none)
trace_subject: (none)
gain_reading: LLM agents with containerized tools completed five proteomics and RNA-seq reanalyses that closely reproduced authors' deposited abundances, supporting automated reuse of published omics data.
gain_evidence: All five reanalyses completed | reproduced the authors' deposited abundances with high per-sample correlation (0.85-0.997)
problem_reading: (none)
problem_evidence: (none)
quick_read: Researchers built an agentic framework in which large language model agents fetch omics studies, extract article metadata, locate and download published data, run containerized quantification pipelines, and synthesize results across studies. It was applied to thousands of PubMed Central articles to catalogue dataset references, and then to five full reanalyses.

The work matters because most published omics data are hard to reuse computationally, since the information needed to reproduce results is scattered across text, supplements, and code. The authors present only a feasibility demonstration, so how well the approach works across the wider literature remains untested and needs prospective evaluation.
limitation: The authors describe the work as a feasibility demonstration, not a validated benchmark of literature-wide performance, and the corpus-scale outputs are not a validated measure of extraction accuracy.
tag: Evidence-backed gain
key_points: The agents used model context protocol servers to expose containerized analysis tools. | Five reanalyses spanned data-dependent and data-independent proteomics and bulk RNA-seq. | Differentially expressed feature fold changes were strongly concordant (Spearman 0.88-0.91), with no direction reversals among features called in both analyses. | A random-effects meta-analysis recovered consistent protein regulation in liver fibrosis.
rundown: The system is designed to address dispersed information: raw data sit in public repositories, but details needed for reproduction are spread across main text, supplementary files and code repositories, and intermediate files such as protein abundance tables are shared irregularly.

Every reanalysis completed, but each involved documented human guidance and workflow accommodations. Remaining differences in significant-feature lists were attributed to threshold placement, tool version, and preprocessing rather than to the underlying quantities.

Beyond reanalysis, the agents identified semantically similar studies, judged whether their data were compatible, and synthesized findings across them. The corpus-scale catalogue is reported as descriptive output, not a validated measure of extraction accuracy.
sources:
- peer_reviewed | PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1014822 | 2026-10-06
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

2fe7c9c535823cc6112bc179160217c47d508837f2c9f73850ea473206f60ce0

previous

0000000000000000000000000000000000000000000000000000000000000000