Overview

FireCloud Method: xavier-genomics/stars_crispr_screen

The STARS CRISPR Screen pipeline takes raw pooled reads (in .fastq format) and does the following:

  1. De-pools samples and generates counts for various genes
  2. If there are multiple complete runs of the sequencing, combine them and treat the separate runs as replicates.
  3. Run edgeR to get differential expression for pairs of conditions.
  4. Combine differential expression results across pairs of conditions.
  5. Perform STARS analysis to identify top gene/comparison combinations.

The eventual output is a file called full_stars_output.tsv that will look something like the following (with many more genes):

full_stars_output.tsv
gene_symbol comparison ranks_of_perturbations_up stars_score_up average_score_up p_value_up fdr_up q_value_up ranks_of_perturbations_down stars_score_down average_score_down p_value_down fdr_down q_value_down best_direction best_p_value best_fdr best_q_value meanLogFC
Rpsa APC_high_vs_APC_low 64082 0.000210041408814 0.000105020704407 0.999952927331568 0.999952927331568 0.999952927331568 32 3.30179612071268 1.65089806035634 0.001583799156615 1 0.845436893203888 down 0.001583799156615 1 0.845436893203888 -9.22135697323133
Mrps14 APC_high_vs_APC_low 64080 0.0002235959564 0.0001117979782 0.999950965970383 1 0.999952927331568 34 3.27546718199033 1.63773359099516 0.001670589389036 1 0.845436893203888 down 0.001670589389036 1 0.845436893203888 -9.20867956964751
Ubc APC_high_vs_APC_low 63860 0.001717184829155 0.000858592414577 0.999607237422771 0.999852371376723 0.999822461991172 254 2.40211238241264 1.20105619120632 0.012126115524174 0.915925925925936 0.899778481012659 down 0.012126115524174 0.915925925925936 0.899778481012659 -8.37619589451366

There are a number of ways to do CRISPR Screen analyses and we may create more pipelines in the future.