User: Xianjun Dong

gravatar for Xianjun Dong
Reputation:
0
Status:
New User
Scholar ID:
Google Scholar Page
Last seen:
2 weeks, 3 days ago
Joined:
4 years, 12 months ago
Email:
x*****************@gmail.com

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

This helps us discourage the inappropriate use of our site.

Posts by Xianjun Dong

<prev • 15 results • page 1 of 2 • next >
0
votes
1
answer
137
views
1
answer
For multiple level variables, different results with or without subsetting
... Hi, I have a simple case, design(dds) = ~ sex + batch + condition, where condition is a factor with three levels (A, B, and C). I tried two different ways of testing differential expressed genes between condition B and C: Method 1: Run DEseq on the whole dataset, and extract comparison between B a ...
deseq2 written 15 months ago by Xianjun Dong0 • updated 15 months ago by James W. MacDonald51k
0
votes
1
answer
373
views
1
answers
Comment: C: Artificial P-value caused by outlier samples
... Last question on this: Would you think setting betaPrior=TRUE in DEseq(), which calculates p-values for the shrunken LFC, can help for outlier cases like this? ...
written 18 months ago by Xianjun Dong0
0
votes
1
answer
373
views
1
answers
Comment: C: Artificial P-value caused by outlier samples
... Thanks, Mike. That's OK. Nothing is perfect. I can still play with the tricks you teach me and filter out those cases, and there are not that many of them. Have a nice weekend! ...
written 18 months ago by Xianjun Dong0
0
votes
1
answer
373
views
1
answers
Comment: C: Artificial P-value caused by outlier samples
... Hi Mike, I really appreciate your time and helps, but I don't think we are on the right question: My true question is why samples with one or two outliers can lead to such a fake "significant" p-value? In the case above (example gene in the DP group), most examples have zero expression and only two ...
written 18 months ago by Xianjun Dong0
0
votes
1
answer
373
views
1
answers
Comment: C: Artificial P-value caused by outlier samples
... Thank you, Mike.  lfcShrink apparently does good job in shrinking the LFC, but seems not affect the p-value. e.g.  > res_after_shrink['ENSG00000257885.1',]                    baseMean log2FoldChange     lfcSE      stat       pvalue         padj                   <numeric>      <numeric ...
written 18 months ago by Xianjun Dong0
0
votes
1
answer
373
views
1
answers
Comment: C: Artificial P-value caused by outlier samples
... Thanks, Michael.  I will definitely add the non-zero counts requirement. But this will work for cases where most samples are at 1 (for example), and two samples are at extremely high (e.g. 100). I am wondering if it's possible (and how) to set the cooksCutoff parameter for case like this? Also, wou ...
written 18 months ago by Xianjun Dong0
0
votes
1
answer
373
views
1
answer
Artificial P-value caused by outlier samples
... HI, I got the following plot for one of the DE genes. The p-value in DP cell type (the most right panel) is 7.50e-07 (FDR = 3.04e-05), with log2(fold-change) of 30. But obviously it's caused by the two outlier samples on the right corner. How would I avoid reporting genes like this? I am not sure t ...
deseq2 written 18 months ago by Xianjun Dong0 • updated 18 months ago by Michael Love26k
0
votes
1
answer
454
views
1
answers
Comment: C: contrast design of DESeq2 for correlated data
... OK. Now I got the all four cell-specific conditions (e.g. "cellA.conditionMS"  "cellB.conditionMS"  "cellC.conditionMS" and "cellD.conditionMS").  Thanks.  Could you please help the other three questions?  Q2: Are the cell-specific comparisons ("cellA.conditionMS"  "cellB.conditionMS"  "cellC.cond ...
written 19 months ago by Xianjun Dong0
0
votes
1
answer
454
views
1
answers
Comment: C: contrast design of DESeq2 for correlated data
... Sorry... I should have read more carefully. Note for the details from https://stat.ethz.ch/R-manual/R-devel/library/stats/html/formula.html: An expression of the form y ~ model is interpreted as a specification that the response y is modelled by a linear predictor specified symbolically by model. ...
written 19 months ago by Xianjun Dong0
0
votes
1
answer
454
views
1
answers
Comment: C: contrast design of DESeq2 for correlated data
... No, I didn't. Here is what I copied from the Rstudio console (sorry the variable name is slightly different): > design(dds) = ~ Batch + Age_bin + Sex + Cell_type + Cell_type * Diagnosis > dds <- DESeq(dds, minReplicatesForReplace=Inf) estimating size factors estimating dispersions gene-wi ...
written 19 months ago by Xianjun Dong0

Latest awards to Xianjun Dong

Popular Question 5.0 years ago, created a question with more than 1,000 views. For How to design formula in DESeq2 when setting blind=FALSE in rlog transformation?

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 193 users visited in the last hour