Data Sheet 1_Harnessing interpretable deep learning to predict resistance in Klebsiella pn

Data Sheet 1_Harnessing interpretable deep learning to predict resistance in Klebsiella pneumoniae.pdf is a 221.5 KB PDF file authored by Nicolas da Matta Freire Araujo and last updated on 2026-05-28. It describes a deep learning model, DeepMDC, which uses whole-genome data to predict antimicrobial susceptibility in Klebsiella pneumoniae for four antibiotics: meropenem, cefepime, ceftazidime, and gentamicin.

Use Cases

• Training interpretable deep learning models for antimicrobial resistance prediction based on genomic data.
• Identifying genomic signatures linked to resistance mechanisms using attention scores.
• Benchmarking predictive performance for Klebsiella pneumoniae against four clinically relevant antibiotics.
• Formulating phenotypic classification as a Multiple Instance Learning problem with whole-genome data.

Strengths

• Model evaluated against four clinically relevant antibiotics (meropenem, cefepime, ceftazidime, and gentamicin).
• Architecture incorporates interpretability through attention mechanisms for biological insight.
• Dataset is openly licensed under CC-BY-4.0.

Limitations

• Dataset size is 221.5 KB, indicating a very limited scope, likely a supplementary document rather than a primary data corpus.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

Nicolas da Matta Freire Araujo via figshare

Collection Method

Likely contains results and methodology from a research study using a deep learning architecture (DeepMDC) on whole-genome data.

Freshness

Last updated 2026-05-28 04:51:25; freshness should be verified.