User: Frederik Ziebell
Frederik Ziebell • 0
- Reputation:
- 0
- Status:
- New User
- Last seen:
- 2 months, 1 week ago
- Joined:
- 1 year, 10 months ago
- Email:
- f********@web.de
Profile information, website and location are not shown for new users.
This helps us discourage the inappropriate use of our site.
Posts by Frederik Ziebell
<prev
• 13 results •
page 1 of 2 •
next >
0
votes
1
answer
154
views
1
answer
... I have a large (bulk) RNA-seq data set with ~1500 samples, i.e. ~30 multiplexing runs (library prep and sequencing) with in total ~500 different conditions in triplicates (conditions are somehow randomized across runs). The ultimate goal is to test all 500 conditions against the wildtype-controls (w ...
written 11 weeks ago by
Frederik Ziebell • 0
• updated
10 weeks ago by
Michael Love ♦ 25k
0
votes
1
answer
117
views
1
answers
... Thank you for the clarification. The actual dataset I have contains many conditions and so lmFit takes about half an hour, that's why I initially viewed it as a bug. Just out of curiosity, what's the super-fast algorithm that only works if there are no NAs? ...
written 11 weeks ago by
Frederik Ziebell • 0
0
votes
1
answer
117
views
1
answer
... Having only a single missing value slows lmFit down by over an order of magnitude:
library("limma")
library("tictoc")
n_genes <- 10^6
sd <- 0.3*sqrt(4/rchisq(n_genes,df=4))
y <- matrix(rnorm(n_genes*6,sd=sd),n_genes,6)
y[1:2,4:6] <- y[1:2,4:6] + 2
d ...
written 11 weeks ago by
Frederik Ziebell • 0
• updated
11 weeks ago by
Gordon Smyth ♦ 39k
0
votes
1
answer
1.1k
views
1
answers
... Hi Mike,
thank you for having a look at my data! I'm also aware of the problem of "almost" confounding, but thought I could get away with it, since most runs share two conditions with other runs. We are currently setting up new runs to re-measure previous conditions and de-confound the design.
How ...
written 17 months ago by
Frederik Ziebell • 0
0
votes
1
answer
1.1k
views
1
answers
... Here is the plot when setting n to the samples in a condition, which makes up a biological group:
Even when I use a more aggressive filter and set n to the number of samples in a run (48), the problem still remains:
...
written 17 months ago by
Frederik Ziebell • 0
0
votes
1
answer
1.1k
views
1
answers
... Hi Mike,
I tried several filtering methods but none seemed to resove the problem. I made the problem smaller by just keeping two genes in the data with similar baseMean and variance. If I use ~condition+run as design, one gene gets a very low dispersion estimate and the other a very high one (upper ...
written 17 months ago by
Frederik Ziebell • 0
0
votes
1
answer
1.1k
views
1
answers
... Hi Mike,
I followed up with your suggestion and figured out the source of the batch effect. It was a different read-length that affected the mapping, since I was using STAR with splice junction overhang equal to read length minus 1. After trimming all reads to the same length, the effect is gone:
...
written 17 months ago by
Frederik Ziebell • 0
0
votes
1
answer
1.1k
views
1
answers
... Hi Mike,
indeed there are batch effects in the data. The way samples are processed is that 48 samples are always multiplexed together. So for these samples, library prep is performed, followed by sequencing and the combination of both is encoded as a "run". Here is a PCA of the vst-transformed coun ...
written 18 months ago by
Frederik Ziebell • 0
0
votes
1
answer
1.1k
views
1
answer
... I have a RNASeq data set with ~650 samples. When I run DESeq2, I get a very poor fit of the mean-dispersion trend:
Here is also a plot of the original trend (without final dispersion estimates), and color coded the number of dispersion iterations:
This is the output of DESeq:
In DESeqDataSe ...
written 18 months ago by
Frederik Ziebell • 0
0
votes
1
answer
579
views
1
answers
... Hi Mike,
thank you for your explanations. The reason I like to use rlog() over vst() is that it achieves almost equal within-sample distributions of counts across all samples, which may be connected to the size factor issue you were mentioning. Do you see potential future improvements of rlog() usi ...
written 21 months ago by
Frederik Ziebell • 0
Latest awards to Frederik Ziebell
No awards yet. Soon to come :-)
Use of this site constitutes acceptance of our User
Agreement
and Privacy
Policy.
Powered by Biostar
version 16.09
Traffic: 314 users visited in the last hour