ALPAR on CAMDA (Critical Assessment of Massive Data Analysis)

CAMDA 2025 - Third best model

Using the ALPAR pipeline, we performed analysis of 5346 bacterial strains from nine species given as the training set by the organizers of the CAMDA anti-microbial resistance prediction challenge 2025, to predict AMR status of 5353 bacterial strains from nine different species (Acinetobacter baumannii, Campylobacter jejuni, Escherichia coli, Klebsiella pneumoniae, Neisseria gonorrhoeae, Pseudomonas aeruginosa, Salmonella enterica, Staphylococcus aureus, Streptococcus pneumoniae) held as the private test set in that challenge.

Bacterium MCC Value
Acinetobacter baumannii0.772
Campylobacter jejuni0.931
Escherichia coli0.463
Klebsiella pneumoniae0.561
Neisseria gonorrhoeae0.141
Pseudomonas aeruginosa0.130
Salmonella enterica0.817
Staphylococcus aureus0.948
Streptococcus pneumoniae0.737

CAMDA 2024 - Winner

Using the ALPAR pipeline, we performed analysis of 5615 bacterial strains from six species given as the training set by the organizers of the CAMDA anti-microbial resistance prediction challenge 2024, to predict AMR status of 1820 bacterial strains from seven different species (Campylobacter jejuni, Campylobacter coli, Escherichia coli, Klebsiella pneumoniae, Neisseria gonorrhoeae, Pseudomonas aeruginosa, Salmonella enterica) held as the private test set in that challenge.

Bacterium MCC Value
Campylobacter jejuni0.968
Escherichia coli0.349
Klebsiella pneumoniae0.886
Neisseria gonorrhoeae0.935
Pseudomonas aeruginosa0.538
Salmonella enterica0.706

We predicted the test set using trained models with the random forest algorithm. All models were trained using the ALPAR Automatix pipeline with the options mentioned above, utilizing species-specific references and protein databases. (Campylobacter jejuni model used for both Campylobacter jejuni and Campylobacter coli). Our predictions achieved a F1-score of 83/100, which was the best performance in the leaderboard.