Help

What is ResLit?

ResLit is a database of antimicrobial resistance (AMR) evidence mined automatically from the published scientific literature. It extracts resistance genes and mutations from full-text papers, links every entry to the paper it came from, and cross-references each against three curated reference databases — CARD, ResFinder, and the NCBI Reference Gene Catalog.

ResLit is built for two audiences. Researchers can look up what the literature reports for a gene, mutation, organism, or antibiotic in one place. Curators can find determinants that appear in papers but not yet in reference databases, and check existing curated entries against their cited sources.

Every entry carries an evidence tier that shows how well supported it is (explained below), and a link to its source publication so you can verify it yourself.

Evidence tiers

Every gene and mutation is assigned one of four tiers, based on two independent lines of evidence: whether it is recorded in a reference database (CARD, ResFinder, or the NCBI Reference Gene Catalog), and whether ResLit independently found it in the literature it searched.

Confirmed and Established entries both rest on reference-database curation (or, for some Confirmed entries, direct curator review) — the difference is only whether ResLit’s own literature search also turned them up. Supported and Candidate entries are literature findings with no reference-database backing yet, and are the best starting point for curation.

Browsing genes

The Browse AMR Genes page lists resistance genes extracted from the literature. Use the filters on the left to narrow the list, the search box to find a gene by name, and the Download button to export the current (filtered) table.

What you can do here

• Search for a gene by name (or by allele, in By Allele view).
• Switch between one row per gene (By Gene) and one row per allele (By Allele) with the toggle at the top of the page.
• Filter the list using the sidebar (see filters below).
• Page through results.
• Click any row to open its full detail view.
• Download the filtered results as a table.

Filters (left sidebar)

• Validation Status — restrict to one evidence tier (Confirmed / Established / Supported / Candidate).
• Resistance Mechanism — filter by mechanism of action (e.g. antibiotic inactivation, efflux).
• Antibiotic — show only genes conferring resistance to a chosen antibiotic.
• Encodes — filter by the product the gene encodes.
• Source Database — show genes recorded in a chosen reference database, or in ResLit only.
• Organism — filter by the species the gene was found in.
• PMID — find genes extracted from a specific PubMed paper.
• Country — filter by country of isolate origin.

Column reference (genes table)

Gene detail view

Clicking a gene opens its detail page, which gathers everything ResLit knows about that gene. The header shows the gene’s mechanism, validation method, drug classes, organisms, geographic locations, allele count (if more than one), paper count, source databases, and — if mutations are recorded for this gene — how many.

External links

• NCBI Pathogen Isolates — bacterial genomes carrying this gene.
• NCBI Reference Gene Catalog — known alleles and reference sequences.
• UniProt — protein entries for this gene.
• AMR Rules — whether the gene confers resistance under manually defined rules.

If a gene has more than one recorded allele, an Allele Variants table breaks down papers, databases, drug classes, organisms, countries, years, and sequence/protein accessions per allele.

Papers— every publication ResLit extracted this gene from is listed, each linked to its PubMed record, along with the gene name and allele, encoded product, source database, mechanism, antibiotics it confers resistance to, organisms tested in, and sequence and protein accessions (linked out to the corresponding NCBI records). If the gene also has mutations, you can toggle between gene papers and mutation papers. Visitors can leave comments on the entry below the papers list.

Validation status and history— the Validation Status badge shows the gene’s current evidence tier. Next to it, a View history control shows how the entry has been curated over time — visible to everyone, not just signed-in curators.

Browsing mutations

The Browse AMR Mutations page lists resistance mutations extracted from the literature. It works like the genes page, with filters tailored to mutations. Use the toggle at the top to view mutations as a flat list (All Mutations) or grouped by their target gene (Browse by Gene).

What you can do here

• Search by nucleotide change, gene, protein change, effect, or paper.
• Switch between the flat list and the by-gene grouping.
• Filter using the sidebar.
• Page through results.
• Click any row to open its detail view.
• Download the filtered results as a table.

Filters (left sidebar)

• Validation Status — restrict to one evidence tier.
• Antibiotic — show mutations conferring resistance to a chosen antibiotic.
• Source Database — show mutations recorded in a chosen reference database, or ResLit only.
• Gene Name — filter by the gene carrying the mutation (hidden while in Browse by Gene mode).
• Mutation Type — filter by type (e.g. substitution, promoter).
• PMID — find mutations extracted from a specific paper.
• Country — filter by country of isolate origin.

Column reference (mutations table)

Mutation detail view

Clicking a mutation opens its detail page. The header shows the gene (linked to its own detail page), the protein and/or nucleotide change, the current Validation Status, and which databases record this mutation, next to a View history control — visible to everyone, not just curators. Below that, an overview lists the mechanism, antibiotics affected, organisms, and how many papers describe it.

Paper information— every publication ResLit extracted this mutation from is listed, each linked to PubMed, showing the protein and/or nucleotide change, mutation type, position, organism, antibiotics affected, effect on function, validation method, and source database reported in that paper.

Visitors can leave comments on the entry.

For curators

Signed-in curators have a dedicated review interface for genes and mutations, reached from the Curator Dashboard after logging in. It mirrors the public catalogues but adds per-entry curation controls.

What curators can do

• Review each gene or mutation alongside its extracted fields and current validation status.
• Edit an entry to correct a field.
• Approve or Reject an entry.
• View history to see how an entry has changed over time.
• Jump from a curation entry to its corresponding public browse page (“View full details”).
• Use the same faceted filters as the public catalogues (validation status, source database, organism, antibiotic, and more) to target entries for review.

Edits and status changes update the live record, but every change is also logged — what changed, from what to what, and when — to that entry’s history, visible via View history. That keeps the extraction pipeline auditable against curator judgement even though the record itself shows only the current value.

Candidate-tier entries — determinants found in the literature but absent from all reference databases — are the natural starting point for curation. Filter to the Candidate tier to focus on them.

To request a curator account, apply as a curator.

Downloading data

Both the genes and mutations catalogues can be downloaded as CSV tables using the Download button on each browse page. You can download the filtered results, curated-only entries, or the entire table (including pending entries). The filtered download respects whatever filters and search you have applied, so you can export exactly the subset you need — for example, all Candidate-tier mutations in a chosen organism — rather than the whole table.

Data sources and how ResLit is built

ResLit assembles its corpus from PubMed, screens papers for AMR relevance, retrieves full text where available, and extracts genes and mutations using a large language model pipeline with a verification step. Each entry is then cross-referenced against CARD, ResFinder, and the NCBI Reference Gene Catalog to assign its evidence tier. See About for a fuller walkthrough of the pipeline.

A manuscript describing ResLit’s methodology in full is currently in preparation. The pipeline source code is available on GitHub.

Frequently asked questions