## User: sven.schenk

sven.schenk10
Reputation:
10
Status:
New User
Location:
Vienna
Last seen:
2 days, 16 hours ago
Joined:
1 month ago
Email:
s**********@univie.ac.at

zoologist gone biochemist, transformed to molecular biologist, over time slowly evolving into a bioinformatician.

#### Posts by sven.schenk

<prev • 11 results • page 1 of 2 • next >
1
83
views
1
... ok, that makes sense so nothing that comes out of DESeq (apart from baseMean) can be used. But I was thinking if baseMean is independent so should be SD(baseMean) this would also most directly account for my varioation issue.   ...
written 2 days ago by sven.schenk10
1
83
views
1
... ok, I guessed as much. so that would mean I should generally not filter like this: interestingGenes<-subset(input, abs(lfcSE*sqrt(3)/log2FoldChange)<=1) but rather use use something involving baseMean,but how would I find an unbiased way to define a cut off for my subset? Could I use baseM ...
written 4 days ago by sven.schenk10
1
83
views
1
... Thanks, lets see if it helps. ...
written 4 days ago by sven.schenk10
1
83
views
1
... Hello, I have rather general question concering FDR/adjusted pvalues. I did an RNASeq experiment (treated vs untreated cells), and analysed the data with DESeq. There I found that I had only very few transcripts passing the FDR cut off at 0.1, which I assumed to be because I have rather low effect ...
written 4 days ago by sven.schenk10 • updated 4 days ago by Michael Love18k
1
48
views
1
... Hi thanks for your answer. Yes, I think roast would do exactly what I'd want, but for that I'd have to have replicates which I don't have. I have only the Counts from one enrichment analysis the Counts from the second one which I want to compare to each other. I then found that I could use geneSet ...
written 9 days ago by sven.schenk10
1
48
views
1
... Hi, I am trying to compare two result lists from GOstats. The reason for doing this is that I did a proteomics experiment in which we established a new sample preparation protocol and I now want to kind of show that with this new method we do not enrich/deplete for certain proteins/protein function ...
written 10 days ago by sven.schenk10 • updated 10 days ago by James W. MacDonald46k
2
114
views
2
... hi,   thanks for your answer. i came to the same conclusion, when checking whats happening during building the GOallframe - in this I have something like 4700 Go terms insteadt of 2472 in my input list. so it was indeed a misunderstanding how GOstats works. thanks again for the answer and the ad ...
written 27 days ago by sven.schenk10
2
114
views
2
... hi again, sorry for not using the COMMENT/REPLY button. no, i did not just update GO.db, i updated the whole installation from 3.3.2 to 3.5.0 it wouldn't work otherwise. yes, you're right that it would be expected that i find more non-matching GO terms if i use a newer DB, however, even if i use a ...
written 4 weeks ago by sven.schenk10
2
114
views
2
... hi, thanks for your answer. so i updated my GO.db from 3.3.0 to 3.6.0 and rerun the analysis, however the problem persits (it even gets more pronounced). so i think the problem is the time of annotaion of my transcriptome which was in July 16 and my old GO.db (3.3.0) was form October 16 wich led to ...
written 4 weeks ago by sven.schenk10 • updated 4 weeks ago by Martin Morgan ♦♦ 21k
2
114
views
2
... addendum: my GOStats code would look something like this: >goFrame_gU.new.H = GOFrame(gU_frame.new, organism="Homo sapiens") >goALLFrame.gU.new.H = GOAllFrame(goFrame_gU.new.H) >ids <-c("comp2258926_c0_seq3","comp2258476_c0_seq2","comp2246178_c0_seq1","comp2268190_c1_seq4") >univ ...
written 4 weeks ago by sven.schenk10

#### Latest awards to sven.schenk

No awards yet. Soon to come :-)

Content
Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.