Moderator: Gordon Smyth

gravatar for Gordon Smyth
Gordon Smyth35k
Reputation:
35,420
Status:
Trusted
Location:
Walter and Eliza Hall Institute of Medical Research, Melbourne, Australia
Website:
http://www.statsci.org...
Scholar ID:
Google Scholar Page
Last seen:
4 hours ago
Joined:
15 years, 10 months ago
Email:
s****@wehi.edu.au

Head of Bioinformatics Division at the Walter and Eliza Hall Institute of Medical Research.

My research group created the limma, edgeR, goseq, Rsubread, csaw and diffHic packages.

Posts by Gordon Smyth

<prev • 3,455 results • page 1 of 346 • next >
0
votes
1
answers
14
views
1
answers
Answer: A: Remove "no_feature" and "alignment_not_unique" counts before using EdgeR or Dese
... I think you will find that edgeR and DESeq2 both remove these rows of the htseq output automatically, so you don't have to do it yourself. But htseq isn't Bioconductor software. You might consider using the Bioconductor package Rsubread instead, which has been benchmarked as faster and more accurat ...
written 5 hours ago by Gordon Smyth35k
0
votes
1
answers
62
views
1
answers
Answer: A: Data import with read.delim
... read.delim() is designed to read just one text file, not a whole set of files. Did you mean to use edgeR::readDGE() instead? But, if you're using Rsubread you should never to read in a text file at all. The Rsubread featureCounts() function produces an R object directly. You can save it to disk usi ...
written 2 days ago by Gordon Smyth35k
0
votes
1
answers
101
views
1
answers
Comment: C: Count filter for differential expression
... I wrote the sentence that you quote so, yes, of course I agree with it. The filterByExpr() function that I recommended in my answer above is in fact designed to choose the cutoffs for you automatically according to the same rule described in the F1000R that you cite. The rule does not tell you to us ...
written 8 days ago by Gordon Smyth35k
0
votes
1
answers
101
views
1
answers
Answer: A: Count filter for differential expression
... Yes, you should definitely should do some filtering. Leaving lots of genes in the analysis with very low counts will just make the dispersion estimation trends more difficult for edgeR and limma to estimate, without any compensating advantage. Your ad hoc filtering is not correct however because yo ...
written 8 days ago by Gordon Smyth35k
0
votes
2
answers
70
views
2
answers
Answer: A: download data from GEO Illumina platform
... See Sections 4.6 and 17.3 of the limma User's Guide: http://bioconductor.org/packages/release/bioc/vignettes/limma/inst/doc/usersguide.pdf To read and normalize the Illumina Beadchip data you can use: > library(limma) > x <- read.ilmn("GSE57691_non-normalized_data.txt.gz",probeid="ID_RE ...
written 9 days ago by Gordon Smyth35k
0
votes
2
answers
189
views
2
answers
Comment: C: edgeR paired samples with blocking
... Why do you run filterByExpr() but then not use the result? That doesn't seem to make any sense. The best cpm cutoff depends on the library sizes, and you haven't told us what they are., so I can't tell whether your filter is sensible or not. Also, why do you use ">3" when your minimum group size ...
written 9 days ago by Gordon Smyth35k
0
votes
1
answers
101
views
1
answers
Comment: C: Count filter for differential expression
... When you say, "raw counts > 10", what do you mean exactly? Did you require the sum of counts for each gene to be > 10? Or every count to be > 10? Or some of the counts to be > 10? I'm not quite clear what your question is. Are you asking us to choose between (1) and (2)? The edgeR and ...
written 9 days ago by Gordon Smyth35k
0
votes
1
answers
75
views
1
answers
Answer: A: Focused gene set tests
... In the ROAST paper (Wu et al, 2010) we tested ROAST on very highly right skew data (exponentially distributed) to assess its robustness and found it still performed well. It's hard to imagine how you could get such right skew values from real data however. ...
written 10 days ago by Gordon Smyth35k
0
votes
2
answers
103
views
2
answers
Answer: A: Interpreting (potentially conflicting) mroast results
... There is no conflict here. The roast p-values are large, so the two sets are not even remotely statistically significant. Also the proportions of Up and Down genes are very small, as would be expected for non-significant sets. So the two pieces of information are in agreement.  As James has pointed ...
written 18 days ago by Gordon Smyth35k
0
votes
1
answers
109
views
1
answers
Comment: C: Inconsistent result between limma/voom and DESeq2
... Well, if the CPM for this gene is higher in cancer, then it would seem that the gene is actually upregulated, and not downregulated as you stated in your question. ...
written 19 days ago by Gordon Smyth35k

Latest awards to Gordon Smyth

Appreciated 6 weeks ago, created a post with more than 5 votes. For A: How barcode-plot enrichment is calculated?
Scholar 6 weeks ago, created an answer that has been accepted. For A: How barcode-plot enrichment is calculated?
Scholar 6 weeks ago, created an answer that has been accepted. For A: another contrast question - edgeR + scRNA-seq + timecourse
Good Answer 6 weeks ago, created an answer that was upvoted at least 5 times. For A: Removing continuous covariate effects in limma analysis
Good Answer 6 weeks ago, created an answer that was upvoted at least 5 times. For A: How barcode-plot enrichment is calculated?
Epic Question 6 weeks ago, created a question with more than 10,000 views. For limma moderated t-statistics and B-statistics
Popular Question 6 weeks ago, created a question with more than 1,000 views. For limma: paired + multiple comparisons + technical replication?
Scholar 6 weeks ago, created an answer that has been accepted. For A: ANOVA-like test via treat() in limma
Scholar 11 weeks ago, created an answer that has been accepted. For A: another contrast question - edgeR + scRNA-seq + timecourse
Scholar 11 weeks ago, created an answer that has been accepted. For A: Building linear model (age effect on gene expression)
Popular Question 11 weeks ago, created a question with more than 1,000 views. For Using write.table with output from topTags [was: report a possible bug of edgeR]
Popular Question 11 weeks ago, created a question with more than 1,000 views. For How do I find up and down regulated genes for each contrast in LIMMA?
Popular Question 11 weeks ago, created a question with more than 1,000 views. For rcmd check does not recognize generic function definitions
Scholar 11 weeks ago, created an answer that has been accepted. For A: ANOVA-like test via treat() in limma
Popular Question 11 weeks ago, created a question with more than 1,000 views. For LIMMA - summarization of Illumina probes for same gene
Good Answer 11 weeks ago, created an answer that was upvoted at least 5 times. For A: conceptual question about FDR, FDR adjusted p-value and q-value
Scholar 3 months ago, created an answer that has been accepted. For A: Why different filtering criteria using CPM to filter the RNA-seq count data in e
Teacher 3 months ago, created an answer with at least 3 up-votes. For A: ANOVA-like test via treat() in limma
Scholar 3 months ago, created an answer that has been accepted. For A: How to plot MD plot in limma
Teacher 3 months ago, created an answer with at least 3 up-votes. For A: Analysis of pathways using edgeR
Teacher 3 months ago, created an answer with at least 3 up-votes. For A: Trouble with Tximport for edgeR
Teacher 3 months ago, created an answer with at least 3 up-votes. For A: Why different filtering criteria using CPM to filter the RNA-seq count data in e
Teacher 3 months ago, created an answer with at least 3 up-votes. For A: nestedF in decideTests: an analogue of Anova with post-hoc t-tests?

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 200 users visited in the last hour