User: Aaron Lun

gravatar for Aaron Lun
Aaron Lun21k
Reputation:
21,340
Status:
Trusted
Location:
Cambridge, United Kingdom
Scholar ID:
Google Scholar Page
Last seen:
24 minutes ago
Joined:
4 years, 2 months ago
Email:
i******************************@gmail.com

I am a research associate in the field of computational biology at the Cancer Research UK Cambridge Institute in the United Kingdom. I am the author and maintainer of the csaw, diffHic, InteractionSet, scrancydar, beachmat, DropletUtils, chipseqDB and simpleSingleCell packages; a co-author and co-maintainer of the scater, SingleCellExperiment and iSEE packages; a co-maintainer of the edgeR package; a co-author of the TENxBrainData package; and an occasional contributor to the limma package.

Posts by Aaron Lun

<prev • 2,358 results • page 1 of 236 • next >
0
votes
1
answers
12
views
1
answers
Answer: A: plotPlatePosition and plate_position
... Your question suggests some fundamental misunderstandings about how this works. example_sce2 only has 72 columns, so why are you trying to assign a vector with 96 elements to plate_position in the column data? Moreover, it's incorrect to subset LETTERS - as you've found out, just because you get a v ...
written 2 hours ago by Aaron Lun21k
0
votes
1
answers
40
views
1
answers
Answer: A: edgeR. Known cancer genes are not differentially expressed?
... Your analysis betrays a certain amount of naivete regarding the processing of human patient data. There are many possible confounding factors that need to be considered in these data sets - for example, I would routinely include age, sex and genetic background/ethnicity in my model. This might possi ...
written 2 days ago by Aaron Lun21k
0
votes
1
answers
50
views
1
answers
Answer: A: Problem installing scater
... scater  is still available and installable from Bioconductor. Your choice of what to show from the installation log is unfortunate as you've removed all the relevant error messages. My guess is that your installation directory (/usr/local/lib/R/site-library) is not writeable by you (as an ordinary ...
written 2 days ago by Aaron Lun21k
0
votes
1
answers
94
views
1
answers
Comment: C: Controlling numerical covariates: differential gene expression with limma voom
... No, I meant any smooth relationship (depending on the number of d.f.). See section 9.6.2 of the limma user's guide. ...
written 3 days ago by Aaron Lun21k
0
votes
1
answers
108
views
1
answers
Comment: C: t-statistic from various limma-approaches does not follow the theoretical distri
... Not unexpected, and not a problem. The exact shape of the QQ plot will depend on the distribution of effect sizes and the power of your experimental design. For example, here's a more "straight-line" QQ plot: set.seed(0) all.t <- rt(10000, df=8) all.t[1:3000] <- 2*all.t[1:3000] # DE genes. ...
written 4 days ago by Aaron Lun21k
0
votes
1
answers
66
views
1
answers
Comment: C: Limma Voom duplicateCorrelation Design
... In general, no. The reason we have to use duplicateCorrelation in the first place is because we don't have enough information in the experimental design to estimate the effect of the blocking factor, which means we can't compute corrected expression values that are free of said effect. At least, not ...
written 4 days ago by Aaron Lun21k
0
votes
1
answers
57
views
1
answers
Answer: A: edgeR DE analysis: single group combined factors or GLM with interaction term
... Your two-step suggestion is the classical approach to testing for main effects. However, the use of empirical Bayes shrinkage in edgeR poses a major problem to fitting the second GLM. Any genes with a significant interaction effect in (1) will have an inflated dispersion in (2). This would subsequen ...
written 5 days ago by Aaron Lun21k
0
votes
1
answers
94
views
1
answers
Comment: C: Controlling numerical covariates: differential gene expression with limma voom
... I'm not technically a limma developer, but I hope my reputation is high enough for you. As Mikhael said, your design matrices and contrasts are correct. I will just note that edgeR uses log-link GLMs, which means that your model assumes that log-expression scales linearly with viral load (i.e., raw ...
written 5 days ago by Aaron Lun21k
0
votes
2
answers
166
views
2
answers
Comment: C: edgeR paired samples with blocking
... The MDS plots won't directly respond to any blocking within duplicateCorrelation. In fact, it's impossible for them to do so, as you can't explicitly estimate and remove the batch effects here without also removing the effect of interest. Any difference will be subtle and due to the effect of duplic ...
written 6 days ago by Aaron Lun21k
0
votes
1
answers
108
views
1
answers
Comment: C: t-statistic from various limma-approaches does not follow the theoretical distri
... Oh, ye of little faith. Perhaps my example would have been more demonstrative with some DE genes: # Mostly non-DE genes with t-distributed t-statistics. # 10% DE with extreme t-statistics in each direction. set.seed(0) all.t <- rt(10000, df=8) all.t[1:1000] <- -5 all.t[1001:2000] <- 5 l ...
written 10 days ago by Aaron Lun21k

Latest awards to Aaron Lun

Teacher 5 weeks ago, created an answer with at least 3 up-votes. For A: Filtering for ATAC-seq
Teacher 5 weeks ago, created an answer with at least 3 up-votes. For A: edgeR normalisation factors different between experimental groups
Teacher 5 weeks ago, created an answer with at least 3 up-votes. For A: Building contrasts for combined treatment groups to compare to a control
Teacher 5 weeks ago, created an answer with at least 3 up-votes. For A: applying voom + limma to a block factor design in RNA-seq experiment
Good Answer 5 weeks ago, created an answer that was upvoted at least 5 times. For A: Is Limma's removeBatchEffect() and log2() commutative?
Teacher 5 weeks ago, created an answer with at least 3 up-votes. For A: Representvie gene expression value in one condition with several replicates
Scholar 5 weeks ago, created an answer that has been accepted. For A: How to extract genes with greatest BCV?
Scholar 5 weeks ago, created an answer that has been accepted. For A: Understanding and creating various comparisons with model.matrix in limma regard
Teacher 5 weeks ago, created an answer with at least 3 up-votes. For A: EdgeR - blocking for multiple factors at once - Errors
Scholar 5 weeks ago, created an answer that has been accepted. For A: applying voom + limma to a block factor design in RNA-seq experiment
Scholar 5 weeks ago, created an answer that has been accepted. For A: EdgeR - blocking for multiple factors at once - Errors
Good Answer 5 weeks ago, created an answer that was upvoted at least 5 times. For A: goana limma- extract list of DE genes and genes in the enriched GO terms?
Scholar 5 weeks ago, created an answer that has been accepted. For A: limma barcodeplot(): calculation of enrichment score
Scholar 10 weeks ago, created an answer that has been accepted. For A: limma barcodeplot(): calculation of enrichment score
Scholar 10 weeks ago, created an answer that has been accepted. For A: How to extract genes with greatest BCV?
Appreciated 10 weeks ago, created a post with more than 5 votes. For A: edgeR normalisation factors different between experimental groups
Teacher 10 weeks ago, created an answer with at least 3 up-votes. For A: Filtering for ATAC-seq
Scholar 10 weeks ago, created an answer that has been accepted. For A: applying voom + limma to a block factor design in RNA-seq experiment
Scholar 3 months ago, created an answer that has been accepted. For A: Problem with annotating ENSEMBLE IDs to GENE SYMBOL with AnnotationDBI mapIDs
Scholar 3 months ago, created an answer that has been accepted. For A: how to calculate the logFC value
Scholar 3 months ago, created an answer that has been accepted. For A: Building contrasts for combined treatment groups to compare to a control
Teacher 3 months ago, created an answer with at least 3 up-votes. For A: applying voom + limma to a block factor design in RNA-seq experiment
Teacher 3 months ago, created an answer with at least 3 up-votes. For A: Representvie gene expression value in one condition with several replicates
Teacher 3 months ago, created an answer with at least 3 up-votes. For A: Building contrasts for combined treatment groups to compare to a control

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 379 users visited in the last hour