uCAREChemSuiteCLI

uCAREChemSuiteCLI predicts bacterial resistome profiles for uncharacterized chemical compounds to forecast potential resistance mechanisms against novel antibiotics.

Key Features:

• Predictive models: Implements a dual-model approach with deterministic and stochastic predictive models based on the premise that drugs with similar structures elicit similar resistome profiles; deterministic model (excluding polyketides and polypeptides) achieves accuracy 87%, sensitivity 85%, precision 89%, and the stochastic model achieves accuracy 72%, sensitivity 75%, precision 83%.
• Target organisms: Predicts resistome profiles specifically for Escherichia coli and Pseudomonas aeruginosa.
• Compound classification: Classifies compounds into 19 predefined drug classes for downstream analysis.
• Pairwise alignments and clustering: Performs pairwise chemical alignments and clusters database compounds to assess similarity relationships among chemicals.
• Chemical structure visualization: Provides visualization of chemical structures to support interpretation of structure–resistome relationships.

Scientific Applications:

• Resistance prediction: Anticipates how Escherichia coli and Pseudomonas aeruginosa might develop resistance to novel chemical entities before clinical testing.
• Antibiotic development prioritization: Informs selection and prioritization of candidate compounds for further investigation in antibiotic discovery.
• Risk mitigation: Supports mitigation of resistance emergence by identifying compounds likely to elicit known resistome responses.

Methodology:

Computational methods explicitly include deterministic and stochastic predictive models built on structural similarity among drugs, pairwise chemical alignments, clustering of database compounds, and classification into 19 predefined drug classes to extrapolate known resistome data to novel compounds.