User: bkellman

gravatar for bkellman
bkellman0
Reputation:
0
Status:
New User
Last seen:
2 years, 1 month ago
Joined:
2 years, 11 months ago
Email:
b*******@eng.ucsd.edu

Profile information, website and location are not shown for new users.

This helps us discourage the inappropriate use of our site.

Posts by bkellman

<prev • 17 results • page 2 of 2 • next >
0
votes
3
answers
1.1k
views
3
answers
Comment: C: Coping with a label/class imbalance in DESeq2
... I realize that normal and nbiom are different distributions. That is why I was comparing the log(fpkm) rather than the fpkm. Log(nbiom) approximates a normal distributions so the results should be very similar. But I agree, rather than theorizing just run the numbers. Here is DE with 2:1 case:contr ...
written 2.9 years ago by bkellman0
0
votes
3
answers
1.1k
views
3
answers
Comment: C: Coping with a label/class imbalance in DESeq2
... Yea, that is what I did, I increased my minimum base mean to mean(log(fpkm_gene))<.5. But after increasing the minimum fpkm, I still appear to be somewhat skewed to lower logFC. (log (fpkm) not fpkm, either way, not a ton of expression) ...
written 2.9 years ago by bkellman0
0
votes
3
answers
1.1k
views
3
answers
Comment: C: Coping with a label/class imbalance in DESeq2
... It is more clear without the gene filter I'm currently using. The low base mean genes (mean(log(fpkm))<.5) are much more biased towards negative logFC. Also, I accidentally responded below. ...
written 2.9 years ago by bkellman0
0
votes
3
answers
1.1k
views
3
answers
Answer: A: Coping with a label/class imbalance in DESeq2
... I looked at the 1st 5k genes. It looks like there is decent (not exceptional) correlation between the deseq estimated logFC (deseq_lFC) and the t.test estimated lFC (t_lFC). > range=1:5000 > t_test = apply(fpkm(dds)[range,],1,function(x) t.test(log(x) ~ dds$dx_clean) ) ) > df=data.frame(d ...
written 2.9 years ago by bkellman0
0
votes
3
answers
1.1k
views
3
answers
Comment: C: Coping with a label/class imbalance in DESeq2
... That is the goal of DESeq but the method used to get there, parameterizing a regression, uses (it should) an error based method. This means it should be sensitive to class imbalance. At the very least, it is weird that I have tons of low logFC upregulation and yet almost exclusively high logFC downr ...
written 2.9 years ago by bkellman0
0
votes
3
answers
1.1k
views
3
answers
Comment: C: Coping with a label/class imbalance in DESeq2
... As far as I understand, coping with sample accuracy is only one use of regression penalties/weights. It may also be used to cope with a class imbalance (combating-imbalance). The issue is that when the parameters are being learned, the over-represented group will drive the learning more since learni ...
written 2.9 years ago by bkellman0
2
votes
3
answers
1.1k
views
3
answers
Coping with a label/class imbalance in DESeq2
... I'm trying to do differential expression on label imbalanced data; my case:control ratio is 2:1. I know that regressions are at the core of DESeq2 machinery and I know regressions have internal machinery for coping with such imbalances. Specifically, you can down-weight observations from the over-re ...
rnaseq regression deseq2 R written 2.9 years ago by bkellman0 • updated 2.9 years ago by Michael Love25k

Latest awards to bkellman

Popular Question 2.1 years ago, created a question with more than 1,000 views. For Coping with a label/class imbalance in DESeq2
Popular Question 2.1 years ago, created a question with more than 1,000 views. For Estimating group-specific dispersion in DESeq2

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 89 users visited in the last hour