1  Datasets

Datasets provideb by CAMDA2025:

• training set: CAMDA training set (training_metadata_db).

  • genus – Bacterial genus assigned to the sample.
  • species – Reported species name; may include annotation inconsistencies based on accession.
  • accession – Unique sample identifier from the original database.
  • phenotype – Reported antimicrobial susceptibility phenotype (e.g., susceptible, intermediate, resistant).
  • antibiotic – Antibiotic used to assess susceptibility.
  • measurement_sign – Comparison symbol used in MIC reporting (e.g., =, <, >).
  • measurement_value – Numeric value of the Minimum Inhibitory Concentration (MIC).
  • measurement_unit – Unit of measurement for MIC (commonly µg/mL).
  • laboratory_typing_method – Method used for strain typing (e.g., sequencing, PCR).
  • laboratory_typing_platform – Instrument or platform used (e.g., Illumina, Nanopore).
  • testing_standard – Testing guideline followed (e.g., CLSI, EUCAST).
  • testing_standard_year – Year of the applied testing standard.
  • publication – Identifier for the original study or source publication.
  • isolation_source – Biological or environmental source from which the sample was isolated (e.g., blood, urine, soil).
  • isolation_country – Country where the sample was collected.
  • collection_date – Date the sample was collected or reported.

• test set: CAMDA test set (test_metadata_db).

  • genus – Bacterial genus assigned to the sample.
  • species – Reported species name; may include inconsistencies depending on the data source.
  • accession – Unique identifier for the isolate or sample.
  • phenotype and measurement_value– In this dataset, represented as "?", indicating that it is an unknown value to be predicted using ML/DL models as part of the CAMDA challenge.
  • antibiotic – Antibiotic tested against the isolate.