User: osa.zuta

gravatar for osa.zuta
osa.zuta40
Reputation:
40
Status:
New User
Location:
United States
Last seen:
1 year, 8 months ago
Joined:
2 years, 3 months ago
Email:
o*******@gmail.com

Posts by osa.zuta

<prev • 9 results • page 1 of 1 • next >
1
vote
0
answers
341
views
0
answers
visualizing top differentially spliced genes
... Hello,  I am using voom/limma to detect differential splicing via diffSplice, with visualization via plotSplice. I am trying to figure out how to make a heatmap of relative exon usage for my top hits, analagous to a heatmap of expression which illustrates the top hits of ordinary differential expre ...
alternative splicing limma voom written 21 months ago by osa.zuta40
0
votes
6
answers
1.5k
views
6
answers
Comment: C: modeling zero-dominated RNA-seq with voom/limma and hurdle models (pscl)
... hi gordon,  thanks! I'll make a new question for edgeR, but I do have one in trying to compare AveLogCPM computed with the QL DEG calls of edgeR with AveExpr (topTable) from a naive application of voom on sc RNA data. Like I said at the beginning, there is good overlap, a reasonable MA plot from pl ...
written 2.2 years ago by osa.zuta40 • updated 21 months ago by Gordon Smyth32k
0
votes
6
answers
1.5k
views
6
answers
Comment: C: modeling zero-dominated RNA-seq with voom/limma and hurdle models (pscl)
... Hi Aaron,  I'm modifying an earlier query about AveLogCPM reported by estimateDisp -- I was confused about how aveLogCPM and the lower-level mglmOneGroup function worked and whether or not they are appropriate for scRNA data, which are zero dominated. The computed average abundances for most genes ...
written 2.2 years ago by osa.zuta40 • updated 2.2 years ago by Gordon Smyth32k
0
votes
6
answers
1.5k
views
6
answers
Comment: C: modeling zero-dominated RNA-seq with voom/limma and hurdle models (pscl)
... logFC verified or rejected by qPCR ...
written 2.2 years ago by osa.zuta40
0
votes
6
answers
1.5k
views
6
answers
Answer: A: modeling zero-dominated RNA-seq with voom/limma and hurdle models (pscl)
... I realize my comment was confusing and revised it as you answered -- all I meant is that the QL estimates inform which logFC are significant.  ( All I mean with the log was that for NB counts c, var(c) ~ mu + d*mu^2, where mu  = average count over biological replicates, d dispersion. Then we can wr ...
written 2.2 years ago by osa.zuta40
0
votes
6
answers
1.5k
views
6
answers
Answer: A: modeling zero-dominated RNA-seq with voom/limma and hurdle models (pscl)
...  I guess I don't completely understand your comment so I'll have to think about it  some more. It's difficult to understand the regime of very high tagwise and common dispersion -- that's part of the problem.  ...
written 2.2 years ago by osa.zuta40
0
votes
6
answers
1.5k
views
6
answers
Answer: A: modeling zero-dominated RNA-seq with voom/limma and hurdle models (pscl)
... Yes, I had set robust = T in glmQLFit already.  In fact, I find it difficult to understand why this setting doesn't affect the MA plot all that much. Would it be too much trouble for you to post a typical MA plot you obtain from scRNA data so I can compare with mine? I'd attach mine but don't know h ...
written 2.2 years ago by osa.zuta40
0
votes
6
answers
1.5k
views
6
answers
Answer: A: modeling zero-dominated RNA-seq with voom/limma and hurdle models (pscl)
... Thanks alot Aaron and Gordon. I've actually tried the QL functions you cite and have found reasonable counterexamples to both the voom and QL DEG calls, i.e. cases where both the QL edgeR calls and voom/limma calls are clearly wrong, but also very high overlap in DEG calls at most FDR cutoffs (thoug ...
written 2.2 years ago by osa.zuta40
11
votes
6
answers
1.5k
views
7 follow
6
answers
modeling zero-dominated RNA-seq with voom/limma and hurdle models (pscl)
... Hello,  Two questions a) can one reasonably use voom/limma to handle single cell RNA-seq data, which are zero-dominated? Naive application seems to yield reasonable results.  b) it seems more appropriate to try to model the process as a hurdle (logit zero  counts, negative binomial finite counts) ...
limma voom single cell hurdle written 2.2 years ago by osa.zuta40

Latest awards to osa.zuta

Popular Question 20 months ago, created a question with more than 1,000 views. For modeling zero-dominated RNA-seq with voom/limma and hurdle models (pscl)

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 131 users visited in the last hour