User: Aaron Lun

gravatar for Aaron Lun
Aaron Lun13k
Reputation:
13,160
Status:
Trusted
Location:
Cambridge, United Kingdom
Scholar ID:
Google Scholar Page
Last seen:
20 minutes ago
Joined:
2 years, 5 months ago
Email:
a***@wehi.edu.au

I am a research associate working in computational biology at the Cancer Research UK Cambridge Institute in the United Kingdom. I am the author and maintainer of the csaw, diffHic, InteractionSet and scran packages, a contributor and co-maintainer for the edgeR package, and an occasional contributor to the limma and scater packages.

Posts by Aaron Lun

<prev • 1,427 results • page 1 of 143 • next >
0
votes
1
answers
33
views
1
answers
Answer: A: Using the ERCC spike as a normalization factor for a custom algorithm
... It seems unwise to use ERCCs for scaling normalisation in bulk RNA-seq experiments, see http://dx.doi.org/10.1038/nbt.2931. I suspect that this is because the definition of "an equal amount" is not precise in bulk settings. For example, it might be an equal quantity of spike-in RNA per tube; or per ...
written 44 minutes ago by Aaron Lun13k
0
votes
1
answers
293
views
1
answers
Comment: C: Disagreement between TMM (edgeR) and RLE (DESeq) normalisation for single-cell d
... Unfortunately, this is where theoretical elegance runs into practical problems with interpretation. The issue, as I discuss in the F1000Research link above, is that spike-in normalisation preserves differences in total RNA content between cells, whereas normalisation by endogenous gene counts does n ...
written 19 hours ago by Aaron Lun13k
0
votes
1
answers
39
views
1
answers
Answer: A: (batch) corrections of RNA-seq data: integrating LIMMA and SVA
... 1) You don't mention the type of error you're getting, but if you look at the documentation for svaseq, it says that the supplied data matrix dat will be log-transformed after adding constant. Clearly, giving the function log-values to be log-transformed again would be silly. 2) Well, it's probably ...
written 19 hours ago by Aaron Lun13k
0
votes
1
answers
68
views
1
answers
Comment: C: How to normalize chromatin and RNA-Seq data together ?
... Indeed, even just trying to fit the same GLM to counts from different data types is a bit of a stretch. For starters, they will have different modes of variability, so trying to estimate a single dispersion value for each "gene" will be inappropriate, let alone trying to fit a mean-dispersion trend. ...
written 20 hours ago by Aaron Lun13k
0
votes
1
answers
103
views
1
answers
Comment: C: Strange output from goana()
... Based on your response to Gordon, the proportion of up-regulated genes in the first set is 124/677 = 0.18. The proportion of down-regulated genes is 113/990 = 0.11. So it's actually a fairly big difference in the proportions, which explains the difference in the p-values corresponding to each direct ...
written 1 day ago by Aaron Lun13k
0
votes
1
answers
103
views
1
answers
Answer: A: Strange output from goana()
... The extra piece of information that determines the p-value calculation is the total number of up- or down-regulated genes. The numbers reported in Up and Down are the number of genes that are up- or down-regulated, respectively, and in the gene set corresponding to the row. Even if these numbers are ...
written 2 days ago by Aaron Lun13k
0
votes
1
answers
43
views
1
answers
Answer: A: Run windowCounts only on annotated regions?
... No. The closest you can get is to run windowCounts on the whole genome (or selected chromosomes), and then to apply findOverlaps to select windows that overlap your regions of interest. Depending on what you want to use the counts for, it may be more appropriate to construct bin/window intervals you ...
written 5 days ago by Aaron Lun13k
0
votes
1
answers
79
views
1
answers
Answer: A: edgeR: time series analysis
... As long as your model has non-zero residual d.f., you can estimate a dispersion for each gene. With time series, the general assumption is that expression follows some smooth trend with respect to time - deviations from that trend can be used for dispersion estimation. Obviously, the more residual d ...
written 6 days ago by Aaron Lun13k
0
votes
2
answers
79
views
2
answers
Answer: A: apply edgeR TMM normalization without converting data format into 'DGEList'
... Your second approach is closer to what edgeR does internally. Normalization is performed by conceptually dividing the counts by the effective library size. The effective library size is, in turn, defined as the product of the normalization factor and the library size for each sample, to account for ...
written 6 days ago by Aaron Lun13k
0
votes
2
answers
1.0k
views
2
answers
Comment: C: edgeR TMM values
... Best to start a new post for a new question. ...
written 6 days ago by Aaron Lun13k

Latest awards to Aaron Lun

Scholar 18 hours ago, created an answer that has been accepted. For A: (batch) corrections of RNA-seq data: integrating LIMMA and SVA
Teacher 18 hours ago, created an answer with at least 3 up-votes. For A: Limma-voom contrast confusion
Scholar 5 days ago, created an answer that has been accepted. For A: Run windowCounts only on annotated regions?
Teacher 6 days ago, created an answer with at least 3 up-votes. For A: Limma-voom contrast confusion
Scholar 6 days ago, created an answer that has been accepted. For A: Can featurecounts count number of mapped reads rapidly in arbitrary regions?
Teacher 11 days ago, created an answer with at least 3 up-votes. For A: Limma-voom contrast confusion
Teacher 19 days ago, created an answer with at least 3 up-votes. For A: Discrepancy in the output of decideTests (total not equal to sum of up/down gene
Teacher 23 days ago, created an answer with at least 3 up-votes. For A: Discrepancy in the output of decideTests (total not equal to sum of up/down gene
Scholar 23 days ago, created an answer that has been accepted. For A: Can featurecounts count number of mapped reads rapidly in arbitrary regions?
Teacher 25 days ago, created an answer with at least 3 up-votes. For A: Discrepancy in the output of decideTests (total not equal to sum of up/down gene
Scholar 25 days ago, created an answer that has been accepted. For A: Can featurecounts count number of mapped reads rapidly in arbitrary regions?
Appreciated 27 days ago, created a post with more than 5 votes. For A: Combining newer/older RNAseq data, batch correcting
Good Answer 27 days ago, created an answer that was upvoted at least 5 times. For A: limma Time Course Experiment with many time points
Scholar 28 days ago, created an answer that has been accepted. For A: Can featurecounts count number of mapped reads rapidly in arbitrary regions?
Teacher 28 days ago, created an answer with at least 3 up-votes. For A: Discrepancy in the output of decideTests (total not equal to sum of up/down gene
Teacher 10 months ago, created an answer with at least 3 up-votes. For A: deseq2, edger different number of replicates and extraction kit
Teacher 10 months ago, created an answer with at least 3 up-votes. For A: Possible implementation of multifactorial contrasts in limma regarding a microar
Teacher 10 months ago, created an answer with at least 3 up-votes. For A: duplicateCorrelation and Pearson Correlation
Teacher 10 months ago, created an answer with at least 3 up-votes. For A: limma Time Course Experiment with many time points
Teacher 10 months ago, created an answer with at least 3 up-votes. For A: Rsubread featureCounts with strand information
Teacher 10 months ago, created an answer with at least 3 up-votes. For A: Limma un moderated t-test
Teacher 10 months ago, created an answer with at least 3 up-votes. For A: Csaw : filter windows based on matched negative control
Teacher 10 months ago, created an answer with at least 3 up-votes. For A: Differential expression analysis: comparison of DESeq2 and edgeR robust results
Teacher 10 months ago, created an answer with at least 3 up-votes. For A: voom with combat
Teacher 10 months ago, created an answer with at least 3 up-votes. For A: Design and contrast question limma (additive or nested or duplicateCorrelation()

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 103 users visited in the last hour