Recount-style output for Pandey's data

This document describes the files at /dcl01/leek/data/recount_pandey created by Leonardo Collado Torres. Please get in touch with him if you have any questions regarding these files.

Recount-style output

The scripts main_job.sh and recount_pandey.sh create objects similar to the ones from the recount project using scripts available at recount-website/recount-prep. The files that get created are:

counts_exon.tsv.gz
counts_gene.tsv.gz
rse_exon.Rdata
rse_gene.Rdata
rse_jx.Rdata
bw/
    mean.bw
    
## Log files
logs/
    pandey_recount*

Which exon-exon junctions overlap Intropolis v2

The scripts query_intropolis.sh, query_intropolis.R and classify_intropolis.R are all used together in order to check which exon-exon junctions from the rse_jx.Rdata file are present in Intropolis v2. The main output of this script is rse_jx_with_intropolis.Rdata which is just like rse_jx.Rdata but includes the logical column is_intropolis labeling if the exon-exon junction was found in Intropolis v2.

The full list of outputs from these scripts are:

## Temporary files used by the scripts
jx_in_intropolis.tsv.gz
jx.tsv.gz

## Log files
logs/
    pandey_intropolis*

## Final output
rse_jx_with_intropolis.Rdata

Exploring junction results

The R script characterize_intropolis.R explores the file rse_jx_with_intropolis.Rdata and creates a series of exploratory plots saved in three different PDF files. These are:

exploratory_plots.pdf
maximum_coverage_jx_Intropolis_not_annotated_UCSC.pdf
maximum_coverage_jx_new.pdf

## Log file:
logs/
    characterize_intropolis_log.txt

Merging by cell

The R script merge_cells.R uses the information from CellMap_codes.csv to merge the technical replicates. It creates the rse_with_cell directory which contains the RangedSummarizedExperiment objects with the cell information but prior to merging by cell type. These files could be useful for some quality control or other analyses. Then merge_cells.R creates the output directory rse_merged with the RSE objects that have merged the information from the cells (that is, 34 columns instead of 258).

The files in rse_merged are:

rse_merged/
    rse_exon.Rdata
    rse_gene.Rdata
    rse_jx.Rdata
    
## Log file:
logs/
    merge_cells_log.txt

Exploring junction results by cells

Then, the script rse_merged/characterize_intropolis_merged.R used the information from rse_merged/rse_jx.Rdata to create another set of exploratory plots (analogous to the ones created previously) to explore the exon-exon junctions found against the number of cells (instead of samples or technical replicates). It creates the PDF files:

rse_merge/
    exploratory_plots_merged.pdf
    maximum_coverage_jx_Intropolis_not_annotated_UCSC_merged.pdf
    maximum_coverage_jx_new_merged.pdf

## Log file:
rse_merged/
    logs/
        characterize_intropolis_merged_log.txt

Custom coverage matrix

The function in the file custom_cov_matrix.R is similar to recount::coverage_matrix() but was changed so that it would work with data that has not been added to the recount resource yet. It takes two new arguments: rse (easiest if you load rse_gene.Rdata) and bigwig_path which in this case is "/dcl01/leek/data/sunghee_analysis/processed/coverage_bigwigs". Check the examples section to see how to run this function.