Postgap analysis needed as input for pipeline

faramer86 / GPrior

GPrior is a Gene Prioritization tool that uses PU learning to prioritize candidate genes based on their similarity with a set of known positive examples. As a result it returns a table with probabilities for each gene.

GNU General Public License v3.0

8 stars 1 forks source link

Postgap analysis needed as input for pipeline #6

Open mohammedmsk opened 2 years ago

mohammedmsk commented 2 years ago

Hello Team,

I am trying to run POSTGAP on my GWAS summary statistics before running Gprior as mentioned in the pipeline.

The issue is that POSTGAP is extremely slow and is not finishing even after running it for days.

Could you please elaborate how to best run POSTGAP on GWAS summary statistics in an efficient way?

Thanks, -Mohammed.

faramer86 commented 2 years ago

Hello Mohammed,

Thank you for your question. Yes, if you run POSTGAP on your local machine, it can be really time-consuming. Thus, if you do have access to server, it would be advisable to use the latter. Otherwise, try to launch summary statistics per chromosome. As alternative, you can try --rsID argument and run POSTGAP per varaint (see their documentation). However, the latter will skip the step in POSTGAP pipeline, when LD cluster collapses. In eitherway, be ready that it will aquire ~week to run one GWAS summary statistics. We are aware of this problem.

mohammedmsk commented 2 years ago

Thanks for your quick reply NK,

I am running it on a server and its still taking a lot of time, main problem is the sequential REST API calls take a lot of time, wish those could be done in parallel
I am trying to run it per chromosome and its still pretty slow for larger chromosomes
The LD clustering step is the most important bit I think which I do not want to skip, so running with --rsID argument is not an option
To be honest a program running for a week is just not optimal, specially when it is not even doing anything which is computationally expensive, I feel its just badly designed.

Would it be possible to please share the postgap output file (And other related files) that you generated for your Schizophrenia example in the paper? That is the main GWAS I am working on.

Thanks, -Mohammed.

faramer86 commented 2 years ago

Hi Mohammed,

It is true, unfortunately, POSTGAP REST API requests take alot of time. All in all, launching POSTGAP is indead very time-consuming step. If you have any other gene-level table you can use it instead of merged POSTGAP results. We are aware of aforementioned problem, thus if there are any other alternatives that can give you gene-level table with annotations, you can freely use it as an input for GPrior.

I need to have some consultations regarding the schizophrenia results. I will answer on your mailbox within several days.