CARD k-mers: Unmasking the pathogen hosts and genomic contexts of antimicrobial resistance genes in metagenomic sequences
Bibliographic record
Abstract
ABSTRACT Antimicrobial resistance (AMR) is a global health crisis requiring rapid surveillance across human, agricultural, and environmental systems. A major challenge during outbreaks is not only detecting antimicrobial resistance genes (ARGs), but also unmasking their pathogen hosts and genomic context, as ARGs alone do not fully capture AMR risk. Pathogen identification is often essential for guiding effective treatment. While culture-based methods remain the diagnostic gold standard, they are slow and sometimes impractical. Faster metagenomic (mNGS) tools typically detect either ARGs, taxonomy, or genomic context, but rarely all three, resulting in fragmented surveillance. Existing k-mer classifiers like Kraken2 and CLARK, designed for general taxonomy, often perform poorly on AMR-specific sequences. We introduce CARD k-mers, the first tool built to jointly predict species-level taxonomy and genomic context (plasmid vs. chromosome) for ARGs in short metagenomic reads. Integrated with the Comprehensive Antibiotic Resistance Database (CARD), CARD k-mers enables rapid, context-aware assignment of ARGs to their likely pathogen and genomic element origin. In benchmarking with 103,456 in- silico pathogen-specific AMR alleles, CARD k-mers outperformed Kraken2 and CLARK by 10.85% and 15.2%, respectively, and correctly classified the genomic context of 4,590 chromosome- and 176 plasmid-specific ARGs. The tool operates at speeds exceeding 675,000 metagenomic reads per minute. By delivering fast, accurate, and context-rich classification of ARGs, CARD k-mers significantly advances untargeted AMR surveillance and is accessible to users with basic command-line experience for use in both clinical and environmental pipelines. CARD k-mers is available at: https://github.com/arpcard/rgi .
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".