gifrop examples

Install gifrop

Make sure you have some kinda of conda installation completed. I like miniconda.
https://www.docs.conda.io/en/latest/miniconda.html

Once you have miniconda install gifrop.

From the command line:

conda create -n gifrop
conda activate gifrop
conda install -c conda-forge -c bioconda -c defaults gifrop

# Test that gifrop was installed correctly

gifrop --help
pan_pipe --help

Clone test repo

If you want to follow along you can clone the repo I’ll be working from.

git clone https://www.github.com/jtrachsel/gifrop_examples.git  
cd gifrop_examples

First Test

Two strains of Salmonella enterica
- LT2: famous reference strain
- USDA15WA-1: Isolate under investigation in the Bearson Lab

Two genomic islands of interest here. 80kb metal tolerance island (SGI4), 28kb multidrug resist module.

For this example I have already annotated the genomes with prokka and generated the pangenome with Roary.

Show gene_presence_absence.csv

cd test1

cd pan

gifrop --get_islands --threads 8  

cd gifrop_out

ls

cd ../../..

Show clustered island info, explain columns. Show gpa_clust.
Show secondary cluster heatmap.

Identified 80kb metal island, 28kb MDR module, phage etc.
- show extracted fasta and gff.

Test 2

Campylobacter draft genomes from an AMR transfer experiment.
1. C. coli donor Tet resistant
2. C. jejuni recipient Tet sensitive
3. C. jejuni result Tet resistant

cd test2

ls

# takes ~7 minutes
pan_pipe --prokka_args '--cpus 2' --roary_args '-p 8' --gifrop_args '--threads 8'

# pan_pipe --prokka_args '--cpus 24' --roary_args '-p 60' --gifrop_args '--threads 60'

cd pan

cd gifrop_out

ls

cd ../../../

Show secondary cluster by genome. Show transfered tet resistance island.

Test 3

A messy dataset. Should show some of the main problems with this approach.

7 Salmonella strains.

I precomputed these results for the sake of time.

pan_pipe --prokka_args '--cpus 2 --proteins LT2.gbk' --roary_args '-p 8' --gifrop_args '--threads 8'

# takes ~ 7 minutes

Look at clustered island info, try and find MDR and SGI4.

Look at secondary cluster by genome heatmap.
Look at secondary cluster 132? heatmap, show problems.
Look at clustering graph.

Potential solutions: