Antimicrobial resistance and virulence gene detection in your browser
openpathogen is a free and open-source platform for screening sequencing reads for antimicrobial resistance and virulence genes directly in your browser. No accounts, no queueing, no uploading your reads to some server somewhere.
1Reads
Drop FASTQ files here or
FASTQ, plain or gzipped. For paired-end reads, select R1 and R2 together.
- R1
- R2
Set automatically from the files you add. Change it here if needed.
2Settings
Optional. Adds CABBAGE tested-resistance rates to the report, and lets PointFinder check the species' known resistance mutations.
Trims adapters and drops low-quality reads before alignment.
Defaults: adapter trimming with auto-detection · quality filtering (min Phred 15, max 40% low-quality bases, max 5 Ns) · length filtering (min 15 bp) · poly-G trimming automatic · base correction off. These are fastp's defaults. Turn a group off to skip that step.
Leave blank for auto-detection (paired runs) or the Illumina TruSeq default shown (single-end). Enter 5 to 100 IUPAC bases, for example a Nextera or custom adapter, to use your own sequence.
A read is dropped if more than the given % of bases fall below the Phred score, or it has more than the N limit.
fastp processes the whole sample, and the databases are run on the trimmed reads. The exact fastp command is printed in the Logs and stored in both reports.
Nanopore reads: fastp is made for Illumina data, so it only reports statistics here. Reads are not trimmed or filtered, and the databases see every read.
For samples with human DNA, such as swabs or clinical metagenomes. Reads that match the human pangenome are removed on this device, and the rest can be downloaded.
Identity is measured over the aligned part of a gene and coverage over the gene's full length. A hit that reaches the identity threshold but not the coverage threshold is listed as a partial gene, with the part that is missing when the reads are deep enough to tell.
How does openpathogen work?
Runs in your browser
openpathogen uses fastp and KMA compiled to WebAssembly and provides indexed database files. Because of this, your sequencing data is processed entirely client-side (in your own computer - which means there's no big, bulky servers anywhere!). No need to worry about accounts, queueing, or data compliance. Once you load the website and fetch the databases, you can even plug it off the internet and it'll still work, even on a smartphone ;).
Databases
ResFinder and CARD for acquired resistance genes, VFDB for virulence factors, and PointFinder for the known resistance mutations of your species (E. coli, K. pneumoniae, S. aureus, Salmonella and six more). Each index is downloaded once, checked against a pinned SHA-256 hash and cached in the browser.
Results by drug class
Resistance genes are grouped by the drug classes and antibiotics they affect, from ResFinder's curated table. When the species is known, each gene also shows how often isolates of that species carrying it tested resistant in CABBAGE, a database of 170,000 sequenced and tested isolates.
Frequently asked questions
- Is my sequencing data uploaded anywhere?
- No. FASTQ files are read into browser memory, processed by WebAssembly and discarded when you close the tab. The only downloads are the app, the databases and, if you ask for one, a public run from ENA. You can read more about it in the privacy section.
- Which inputs are supported?
- Illumina paired-end, single-end and Oxford Nanopore reads in FASTQ, plain or gzipped. You can also fetch any public run by its SRA or ENA accession. Files up to 1 GB each are handled in memory.
- What do the results contain?
- Resistance genes by drug class and names of the antibiotics they affect, the resistance mutations found in the chromosome (when PointFinder covers your species), a list of the classes with no acquired gene, and a summary of the virulence factors. You can print it or save it as a PDF. Ofcourse, you can also view the full results in the browser, including a gene table per database, a viewer per gene with coverage plot, consensus sequence, alignment and variants, the fastp reports, and a ZIP with the full KMA output. Every gene hit also downloads in the PHA4GE hAMRonization format, which combines with the results of other AMR tools.
- Can openpathogen remove human reads?
- Yes. Tick Remove human reads and Deacon removes reads that match the human pangenome before the databases run, on your device. Its 239 MB index downloads once. The report says how many reads were removed, and the remaining reads can be downloaded. Deacon removes most human reads but not every one, so check the rules that apply before sharing them.
- How do I cite openpathogen?
- A short communication is in preparation. Until then, cite KMA, fastp and CABBAGE, Deacon if you removed human reads, and the databases you used.