PhageHost is a set of python scripts that predict phage-host interactions for
E. coli, K. pneumoniae and A. baumannii phages, using supervised machine learning models. The models were built from a dataset containing 252 features and 23 987 entries with balanced outputs of 'Yes' and 'No'. The positive cases of interaction predicted are described in the file "NCBI_Phage_Bacteria_Data.csv", contained within this tool, while the negative were randomly assigned by pairing phages with bacteria of different species. The prediction resorts to complete host proteome and to phage tail proteins, that are inferred within the tool. This inference is made with a locally created database of phage protein functions, available in the file "phagesProteins.json". Unknown proteins are predicted against this database. To help with this prediction, the use of InterProScan is made optional. |
hg clone https://toolshed.g2.bx.psu.edu/repos/pedro_araujo/phage_host
Name | Description | Version | Minimum Galaxy Version |
---|---|---|---|
prediction of phage-bacteria interactions | 0.1.0 | 16.01 |