/splicing_project_moreau

Collab

Primary LanguagePython

splicing_project_moreau

Scripts available here are in Python3. It's not required but advised to install Conda if python3 is not set up on your computer. It will make things easier then for installing tools or switching to older python version if needed.

Look to INSTALL, every steps are documented. First you need to install Whippet here

Here you will find the project manager here Just in case if we want a clean follow-up of what we are doing.

1. Create json config for each fastq pair

Each time, you need to change python3 path accordingly to your home user path.

It creates all the json files used by the script. The json files are used to define path to input,output directories and name of each samples. The script will create a file called listing.json.per.sample.txt with all links to json configs to process sequencially in the next step.

     /home/jean-philippe.villemin/anaconda3/bin/python3 /home/jean-philippe.villemin/splicing_project_moreau/src/createJsonConfForPsi.py \
     -l /home/jean-philippe.villemin/samples.txt \
     -i /eqmoreaux/data/global_data/MMC/RNAseq/FASTQ/ \
     -o /home/jean-philippe.villemin/moreau_splicing/ \
     -s /home/jean-philippe.villemin/splicing_project_moreau/bash/whippet_filter_eventType_wrapped_for_psiOnly.sh

The output directory is set to moreau_splicing. Inside you will find config dir with json files and output dir where splicing results will be written.

2. Process sequentially for splicing analysis

	cat listing.json.per.sample.txt| xargs -n 1  -I %  wrapper.sh %

It will read each line of listing.txt to access each json config. It's a loop in fact...

Wrapper.sh calls the main script as follows using the config path received :

VERY IMPORTANT :

Now you need to set paramerers -j, -i and -r. -j : is the path to julia executable -i : path to whippet jls index -r : path to dir where you find annotSymbol.R

So you need to modify accordingly to your own conf.

You can test for one sample if it works.

nohup /home/jean-philippe.villemin/anaconda3/bin/python3 /home/jean-philippe.villemin/splicing_project_moreau/src/splicingWhippetPSI.py \ --config /home/jean-philippe.villemin/moreau_splicing/config/E10061.json  -i /home/jean-philippe.villemin/index_whippet_human.jls \ 
 -r /home/jean-philippe.villemin/splicing_project_moreau/Rscript/ -j /home/jean-philippe.villemin/julia-d55cadc350/bin/julia &

Then for all

nohup cat listing.json.per.sample.test.txt | xargs -n 1  -I %  ../splicing_project_moreau/bash/wrapper.sh %  &

Read whippet output to get an idea of what we have at the end.

Here I show you a screenshot of what we have in the directory output for one experiment treated.

alt text

3. Create PSI Matrice by concatening all Samples

Take any file.CE.psiannoted.csv and apply the following command to regenerate it properly from your dataset.

/usr/bin/gawk -F ","  'BEGIN {OFS="\t";} \
 {  if ( match($4, "^(chr.*):([0-9]+)-([0-9]+)", ary) ) print ary[1],ary[2]-1,ary[3],NR-1,0,$5,$2 ;}' /home/luco/PROJECT/BEAUTY/output/EX173639.CE.psiannoted.csv > CE.whippet.tsv

You should end with something as follows :
alt text

DO THAT FOR ALL KIND OF EVENTS.
You should end with AD.whippet.tsv , AA.whippet.tsv etc...(a total of 8 files for 8 types of event)

This will create a matrice named eventType.output.tsv for the event you will process. Here CE.whippet.tsv. Rename the output.tsv after each run to not rewrite in it.

/home/jean-philippe.villemin/anaconda3/bin/python3 /home/jean-philippe.villemin/splicing_project_moreau/src/prepareDataForHeatmap.py \ -l /home/jean-philippe.villemin/moreau_splicing/ID_2_GROUP.tsv -d /home/jean-philippe.villemin/moreau_splicing/output/ \
 -e /home/jean-philippe.villemin/moreau_splicing/output/CE.whippet.tsv -t CE
  • -t : Specify the type of event you process. This will be used in the output name. Here the output will be eventType.output.tsv.

Examples are given in config dir for the files used by the script. BEAUTY_ID2GROUP.tsv is just a several columns file with nameOfPatient & annotations.

I strongly advise to regenerate CE.whippet.tsv yourself or at least check it correspond to file.CE.psiannoted.csv (same number of lines, orderded the same way...) It depends of the version gtf you used. Exons can be ordered differently, removed or added.

4. Apply filters to this matrice

Now you get the whole matrice, you need to apply fitlters. You can do it by yourself with your own scripts or I provide some tools that will create a output_filtered.tsv file

/home/jean-philippe.villemin/anaconda3/bin/python3 /home/jean-philippe.villemin/splicing_project_moreau/src/filterHeatmap.py  -m /home/jean-philippe.villemin/moreau_splicing/output.tsv -g /home/jean-philippe.villemin/splicing_project_moreau/config/genes.txt -o CE_JP -a 3

This script work with a provided list of genes symbol. So your are not force to use a gene lists. It's optional.

  • -o : let you set name of your output. here output will be called CE_JP.output_filtered.ts
  • -a : It's the number of annotation you used. The number of columns you set in ID_2_GROUP.tsv.

You filtering by event if more than 25 % NA values computed and if variance < 0.05 for the event.(arbitrary , need to improve that)

5. Use Morpheus here to visualise the matrice in a heatmap.

At the end a quick way to visualize your matrice is this tool. Try it, play with it. It's cool.

Several advices : First change color settings (set white for 0 and blue for 1). Use relative color setting. A pply hierarchical clustering per columns & rows. Or you can apply kmeans clustering techniques.

You should end with something as follows : alt text