Tried some more GO-MWU stuff - not great - switching gears

Today I tried some more GO-MWU stuff, but it didn’t really do much. After input at lab meeting, I tried using DAVID first, then GO-MWU… got 1 enriched biological process. Meeting with Steven –> new plan of things to focus on for now (writing, results we have). Details in post.

Earlier this morning

Before lab meeting, I tried GO-MWU, modelled after Zostera-GO_MWU.R from the EIMD FHL 2019 class. Separates the GO enrichemnt into the different categories: biological process (BP), cellular component (CC), and molecular function (MF).

I also tried it with different input files (DESeq 2019 crab data file (all genes, not just DEGs) with continuous measure of significance):

I tested each one by changed the input in this script: analyses/GO-MWU/2019-crab-infection-ALL-GO_MWU.R

The results in terms of the identified number of GO terms at 10% FDR differed depending on the significance value used:
Molecular Function
Using log2fc → 15 GO terms at 10% FDR
Using padj → 4 GO terms at 10% FDR
Using pvalue → 8 GO terms at 10% FDR
Cellular Component
Using lof2fc → 10 GO terms at 10% FDR
Using padf → 5 GO terms at 10% FDR
Using pvalue → 4 GO terms at 10% FDR
Biological Process
Using log2fc → 7 GO terms at 10% FDR
Using padj → 12 GO terms at 10% FDR
Using pvalue → 4 GO terms at 10% FDR

Then, I saved what the results were from using the analyses/GO-MWU/2019-crab-ALL-geneID-log2fc.csv as the input:
Result files

Trees

When I tried with the analyses/GO-MWU/2019-crab-ALL-geneID-padj.csv file, which is the continuous measure of significance used by EIMD and the GO-MWU example script, it gave essentially no results, even though it found a few GO terms that were at 10% FDR:
Result tree: analyses/GO-MWU/2019-crab-infection-GOBP-padj.pdf
The red and blue colors are missing, and it says that the ratios are zeros…

After lab meeting

During lab meeting, a plan for the afternoon was identified:

  1. Put the Uniprot Accession IDs in DAVID to look for enrichment
  2. If enrichment –> go back to GO-MWU and use BP
  3. If no enrichment –> try something else… possibly TopGO

I focused on trying out DAVID.

Script for creating lsits of uniprot Accession IDs here: scripts/042120-2019crab-infection-for-DAVID.Rmd
Basically, I got 2 lists:

  1. Background –> Uniprot Accession IDs from the BLAST output from the entire crab transcriptome
  2. gene list –> Uniprot Accession IDs from the list of 772 DEGs (made by SR) by join-ing the DEG list with the crab transcriptome BLAST output to get the Uniprot Acccession IDs.

I put the gene list Uniprot Accession IDs in DAVID, then the background Uniprot Accession ID list.

Resulting file for GO_BP_ALL from DAVID –> analyses/2019-crab-infection-DAVID-output.txt

I then tried GO-MWU again, using the padjusted values, and the same thing described above happened.

Then Steven and I met to discuss next steps.

SR crab chat

Steven did the same thing in DAVID as I did, but saved the GOTERM_BP_DIRECT list, which I guess is shorter:
/analyses/chart_12EFD.txt

Either way, the top line, which is the enriched, is GO:0044351~macropinocytosis, and the only result.

I’m going to stop playing around with GO-MWU for the next few days and instead focus on writing.

NEXT STEPS

Tasks for the next couple of days:
METHODS

  • what I can for differential expression, enrichment

RESULTS

  • transcriptome results (we have annotated crab transcriptome: maybe may a GOslim plot, list number of genes, etc.)
  • DESEq results (SR performed analysis, but I can write results)
  • enrichement result (GO:0044351~macropinocytosis –> look at the genes involved in this, find some resources on what it is)
  • mortality (have mort. data! make a visual or something - blurb, whatever for now)

DISCUSSION

  • transcriptome results
  • enrichment result –> little blurb
  • mortality blurb
Written on April 21, 2020