Skip to content

andynet/pheri

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

15 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

PHERI: phage host exploration through protein clustering and decision trees

Introduction

PHERI is a tool to predict the phage hosts from the phage sequence in FASTA format. Currently, it is able to predict 50 host genera with high accuracy (~97%). The database of proteins, clustering of proteins and pretrained models were obtained from the preprocessing pipeline and due to the computational complexity of these steps are included in the tool.

Installation

The recommended way to install PHERI is through conda installer. PHERI uses 3rd party dependencies, which can be found in the bioconda and conda-forge channel, and installed automatically with the following command:

conda install -c bioconda -c conda-forge -c andynet pheri

Running PHERI

The input for the PHERI tool should be a single file representing a single phage in the FASTA format. Furthermore, PHERI requires an output directory, where the results will be stored. We included the test set as an example. To run the example, use these commands:

conda create -n test
conda install -c bioconda -c conda-forge -c andynet pheri
pheri ~/miniconda3/envs/test/data/examples/test_phages/phage0000000.fna ./test

Output files

The files with extensions .fna, .faa, .blast, and .vec.tsv are mainly for sanity checks and can be safely deleted.

extension description
.fna The original FASTA file copied.
.faa The genes predicted from FASTA file by Prokka.
.blast The assignments of clusters based on the internal database of genes and their cluster assignments.
.vec.tsv The feature vector used as an input to the classifier.
.res.tsv The main results. The table contains 50 rows for each of the genera currently predicting and the score of the prediction.

Results

The PHERI results on the testing dataset (n=1201).

host TP FP FN TN
Acinetobacter 9 4 2 1186
Aeromonas 6 2 3 1190
Arthrobacter 46 3 2 1150
Bacillus 30 11 5 1155
Brucella 6 0 1 1194
Burkholderia 5 2 2 1192
Campylobacter 11 3 4 1183
Caulobacter 2 0 2 1197
Cellulophaga 1 2 6 1192
Citrobacter 1 4 4 1192
Clostridioides 4 0 1 1196
Clostridium 2 2 3 1194
Corynebacterium 3 0 1 1197
Cronobacter 2 7 3 1189
Cutibacterium 25 0 0 1176
Enterococcus 6 4 4 1187
Erwinia 4 0 5 1192
Escherichia 94 21 33 1053
Flavobacterium 6 4 1 1190
Gordonia 59 2 3 1137
Helicobacter 6 1 0 1194
Klebsiella 5 7 13 1176
Lactobacillus 7 5 4 1185
Lactococcus 43 4 5 1149
Leuconostoc 4 0 0 1197
Listeria 5 2 2 1192
Mannheimia 2 1 1 1197
Microbacterium 22 2 1 1176
Moraxella 7 1 0 1193
Mycobacterium 0 2 1 1198
Mycolicibacterium 321 4 1 875
Paenibacillus 5 0 1 1195
Pectobacterium 1 2 5 1193
Proteus 2 2 1 1196
Pseudoalteromonas 4 4 2 1191
Pseudomonas 45 8 12 1136
Ralstonia 1 1 6 1193
Rhizobium 1 0 2 1198
Rhodococcus 11 1 1 1188
Ruegeria 3 5 0 1193
Salmonella 22 7 20 1152
Shigella 3 6 8 1184
Staphylococcus 36 5 3 1157
Stenotrophomonas 0 0 3 1198
Streptococcus 39 9 1 1152
Streptomyces 27 3 8 1163
Synechococcus 29 3 0 1169
Vibrio 18 3 14 1166
Xanthomonas 2 3 3 1193
Yersinia 2 5 3 1191

Dependencies

The dependencies are automatically installed with the conda installer. This list can help you, when installing from source (not recommended).

  • python >=3.6
  • prokka >=1.14
  • blast
  • scikit-learn =0.18.0
  • scipy <0.19
  • pandas

Citation

If you use our tool, please cite:

Baláž, Andrej, et al.
PHERI-Phage Host Exploration pipeline.
bioRxiv (2020).

About

Prediction of the phage hosts

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published