Pangenomic View on Antibiotic Resistance

Lab notebook the following research:

1. A pan-genome-based machine learning approach for predicting antimicrobial resistance activities of the Escherichia coli strains. HL Her, YW Wu. Bioinformatics 34 (13), i89-i95(2018)
2. PangenomeNet: A pan-genome-based functional network provide mechanistic view on Meropenem Resistome for Escherichia coli strains. Hsuan-Lin Her, Po-Ting Lin, and Yu-Wei Wu. (In preperation)

The code for building [2] are in main/:

follow the number preceding the scripts to reproduce pangenomeNet work. You will need to install many tools. We are working on refactoring the pipeline into snakemake.

File structure:

1. in Drug folder, there are codes for TMACC, CHEBI, all kinds of drug database and chemical descriptor
2. in Genome folder, codes are about pan-genome construction,
   • annotate_parser contain functions to parse annotations from CARD, BLAST, COG, HMMer, pFAM and REsfam
   • context folder contains functions and scripts to construct co-inheritance network
   • genome folder contains scripts to download genome from PATRIC, and calculate genome statistics
   • goldstandard_pair folder contains scripts to validate networks, generate gold-standard gene pairs.
   • pangenome_annotation folder contains scripts to annotation gene clusters from the pan-genome
   • pangenome_build contains scripts used to call CD-hit
   • pangenome_intrinsic_info contains scripts to calculate pan-genome statistics
   • string contains parsing EcoliNet, Mentha and STRING network
3. in MIC folder, codes are mainly related to CLSI re-annotation for resistance phenotypes
4. in Model folder, codes are all about training svm and nb models
5. compare_scoary contains code that compares [1] to Scoary.
6. the Visulazation folder contains all sorts of jupyter notebook, mainly used to plot
7. cleanData.py selects data that fits our need (with MIC instead of disk diffusion, the right species ...etc)
8. in network_analysis contain codes for powerlaw regression