## User: Yunshun Chen

Yunshun Chen490
Reputation:
490
Status:
Trusted
Location:
Australia
Last seen:
9 hours ago
Joined:
7 years, 1 month ago
Email:
y*****@wehi.edu.au

I maintain the edgeR and RnaSeqGeneEdgeRQL packages.

#### Posts by Yunshun Chen

<prev • 36 results • page 1 of 4 • next >
1
51
views
1
... The plotExonJunc function requires the coordinates of the exons/junctions to produce the plot. In particular, it requires the start and end positions, as well as the length and strand information, of each feature to be stored in genes$Start, genes$End, genes$Length and genes$Strand of the ...
written 7 days ago by Yunshun Chen490
1
123
views
1
... To construct a proper contrast, you need to understand the meaning of each coefficient first. In this particular design matrix, the (Intercept) represents the Healthy Patient 1 with no treatment. The coefficient DiseaseDisease1 represents the **difference** between Disease1 and Healthy for Pati ...
written 26 days ago by Yunshun Chen490
2
107
views
2
... You should always refer to the edgeR user's guide. There are several detailed case studies in the user's guide with complete R code and explanation which you can easily follow. We usually recommend the quasi-likelihood approach for differential expression analysis as we found it gives better FDR co ...
written 5 weeks ago by Yunshun Chen490
2
138
views
2
... Is your x a count matrix? If so, have you performed filtering? Lowly expressed genes (especially with zero counts across all samples) can cause problems in dispersion estimation. You can perform filtering using the filterByExpr() function. ...
written 9 weeks ago by Yunshun Chen490
2
608
views
2
... The current Bioconductor release version is 3.3, compatible with R 3.3. It seems you are not using an old version of R. Please update your R to the latest version before you try it again. ...
written 3.2 years ago by Yunshun Chen490
2
1.3k
views
2
... The Simes method was introduced and described in the following paper: R. J. Simes. An improved Bonferroni procedure for multiple tests of significance. Biometrika, 73(3):751~754, 1986. The Simes' method controls the family-wise error rate in the weak sense, i.e., only when all null hypotheses are ...
written 3.2 years ago by Yunshun Chen490
1
465
views
1
Answer: A: Treat p > 1
... Thanks for reporting this. You're right. These p-values are not expected. The glmTreat is designed for two-sided hypothesis tests. The reason for such unexpected behaviors is that the observed fold change (FC) is closer to one FC boundary (threshold) but its likelihood is closer to the likelihood e ...
written 3.3 years ago by Yunshun Chen490
1
2.2k
views
1
Comment: C: Setting up the Design Matrix
... You're right. The two contrasts give you the results for the two comparisons of your interest. It is still possible to test for the main effect of temperature under the interaction model. However, the contrast is not as intuitive as what you would have using the one-way layout. The contrast would b ...
written 3.7 years ago by Yunshun Chen490
1
2.2k
views
1
... Your design looks correct and testing on coef=3 gives you the overall temperature effect. The first three questions can be answered by using this design. However, an interactive model (design) is required in order to test for the interaction effects between factors.   ...
written 3.7 years ago by Yunshun Chen490
1
624
views
1
... Yes, it is a bug in the internal subsetting step. It's been fixed and it will be updated online soon. Note that the dispCoxReidInterpolateTagwise function uses min.row.sum=5 to filter out lowly expressed genes by default. Only genes (or rows) with total sum of counts above 5, which is the last row ...
written 3.8 years ago by Yunshun Chen490

#### Latest awards to Yunshun Chen

Popular Question 3.0 years ago, created a question with more than 1,000 views. For Questiond about the descrepancy between edgeR and limma-voom
Popular Question 3.0 years ago, created a question with more than 1,000 views. For edgeR: regarding Poisson distribution and goodness of
Voter 3.9 years ago, voted more than 100 times.
Scholar 4.0 years ago, created an answer that has been accepted. For A: negative values for trended dispersion with edgeR
Supporter 4.6 years ago, voted at least 25 times.
Teacher 5.0 years ago, created an answer with at least 3 up-votes. For C: Obtaining Differentially expressed gene lists at a Fold change and FDR cut-off

Content
Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.