Closed lcolladotor closed 4 years ago
In your commit messages, you can link them automatically to this GitHub issue (and even close the issue) using some keywords. Check https://help.github.com/en/github/managing-your-work-on-github/linking-a-pull-request-to-an-issue for more info.
As a name & JHPCE location suggestion, use data/drop_flagged_samples_cov_rse.R
and data/drop_flagged_samples_cov_rse.sh
for these two scripts.
Hi,
The goal here is to practice the GitHub management tools, using
sgejobs
, and start editing some of our analysis scripts.The task is to create a new R script that will create a new
cov_rse
object for MDD + BPD. The current object lives atdata/degradation_rse_MDDseq_BiPSeq_BothRegions.Rdata
and is used in several scripts such as in https://github.com/LieberInstitute/goesHyde_mdd_rnaseq/blob/6a722bbfea2041dbb46ac2a66b95c0b8268ba582/wgcna/run_wgcna_combined.R#L17.It currently has the same metadata columns (
colData()
) as therse_gene
filtered object (lowly-expressed genes were removed) as tested in https://github.com/LieberInstitute/goesHyde_mdd_rnaseq/blob/6a722bbfea2041dbb46ac2a66b95c0b8268ba582/wgcna/run_wgcna_combined.R#L20. With thedropped_flagged_samples.R
script we are in the process of re-making therse_gene
objects https://github.com/LieberInstitute/goesHyde_mdd_rnaseq/blob/d12acf804069463b829c8cf9ca3973976580a906/data/drop_flagged_samples.R#L115. We haven't finished that script since we need to add the SNP genotype data (snpPC1, snpPC2, etc) to it, finalize which samples we'll use, then re-compute which genes (and other features) we will filter due to low expression values. But once we do, we need to have a matchingcov_rse
object.So the goal here is to create an R script with it's companion bash script (use
sgejobs
to make this bash script) that will create a newcov_rse
file that is subsetted to the same samples as therse_gene
object from https://github.com/LieberInstitute/goesHyde_mdd_rnaseq/blob/d12acf804069463b829c8cf9ca3973976580a906/data/drop_flagged_samples.R#L115 and has the same phenotype information.Best, Leo