User: Gavin Kelly

gravatar for Gavin Kelly
Gavin Kelly180
Reputation:
180
Status:
Trusted
Location:
United Kingdom / London / Francis Crick Institute
Twitter:
gavinpaulkelly
Scholar ID:
Google Scholar Page
Last seen:
14 hours ago
Joined:
2 years, 6 months ago
Email:
g*************@gmail.com

Statistician in the Bioinformatics & Biostatistics group at the Francis Crick institute

Posts by Gavin Kelly

<prev • 51 results • page 1 of 6 • next >
0
votes
1
answers
95
views
1
answers
Comment: C: How to use subSeq with DNAStringSet and information of GroupedIRanges?
... If 'pos' is a vector telling you the location of the base you're interested in, can you achieve what you want with something along the lines of subseq(data$seq, start=pos, width=1) or, if pos is only defined for certain rows, then you may need to subset data first to be those rows with a defined pos ...
written 1 day ago by Gavin Kelly180
1
vote
0
answers
76
views
0
answers
Job: Bioinformatics Position: Francis Crick Institute, London
... We're looking for an additional person to work in our bioinformatics core facility within the Francis Crick Institute (a biomedical discovery institute dedicated to understanding the fundamental biology underlying health and disease) to support our scientists in the analysis of their data (mainly fo ...
job bioinformatics uk written 6 days ago by Gavin Kelly180
0
votes
1
answers
101
views
1
answers
Answer: A: Multiple comparisons for DE analysis with DESeq2
... Yes, I think your approach of adding a combination factor (design = ~ new_column) is a correct way to carry out the analysis.  It would be possible to achieve something similar if you had a design = ~Cell_component * Type with an interaction, but it wouldn't be as transparent as the approach you've ...
written 14 days ago by Gavin Kelly180
0
votes
1
answers
91
views
1
answers
Answer: A: How to solve out the misunstanding of fake "red" "green" on the heatmap
... I don't think it's necessarily misleading that a bright colour represents a high expression; the heatmap is representing best estimates of fold-changes, rather than significances, and most people interpret them as such.  The ability to detect common directions in expression (even if they have failed ...
written 16 days ago by Gavin Kelly180
0
votes
1
answers
163
views
1
answers
Comment: C: DESeq2 model design with unequal number of time points
... If you have a specific 'shape' of time-profile you want to investigate, then maybe using time as a continuous variable would allow you to compensate for the 'missing' timepoints.  If you kept the time term linear, then you may be able to detect genes that have differential gradient (on a log-like sc ...
written 9 weeks ago by Gavin Kelly180
0
votes
2
answers
219
views
2
answers
Comment: C: strange results from differential methylation analysis using limma
... Assuming the "data_afterpreprocessed.txt" file contains beta-values of normalised samples, in the same order as the samples in the samplesheet file, I can't see anything terribly wrong with the code. Perhaps if you do an plotMD(fit2, column=4) or volcanoplot(fit2, coef=4) you might see what's drivin ...
written 10 weeks ago by Gavin Kelly180
0
votes
1
answers
119
views
1
answers
Comment: C: DESEq2 Paired samples: Baseline Versus each Timepoint
... Minor typo that might confuse a newbie: contrast=c("Timepoint", "Day1", "Baseline") ?     ...
written 10 weeks ago by Gavin Kelly180
0
votes
4
answers
241
views
4
answers
Comment: C: LIMMA paired analysis adjusting for two continuous variables - microarray data
... Don't think there's an issue with the continuous variables being constant within the subjects - does everything work ok if you have a design with of ~bmi + age + class, with a blocking factor of id?   I suppose you could add age:class (and/or bmi:class) interaction if you're worried that a gene's a ...
written 10 weeks ago by Gavin Kelly180
0
votes
4
answers
241
views
4
answers
Answer: A: LIMMA paired analysis adjusting for two continuous variables - microarray data
... I don't think is quite right  - the class:ID term means that a coefficient will be estimated for each unique {ID, class} combination, so you'll get lots of coefficients corresponding to this term (unless you've left ID as a numeric variable - you must convert it to a factor before putting it into th ...
written 10 weeks ago by Gavin Kelly180
0
votes
2
answers
156
views
2
answers
Comment: C: how to plot does response curve for temperature
... Sounds like you're mis-intepreting the advice.  Something along the lines eval(parse(text=TPP:::fctSigmoidTR(deriv=2)), list(Pl=1, a=2, b=3, x=1)) will give you an evaluation of a sigmoidal curve with the parameters that I've put in the list, at the point x=1.  You'll need to extract your fitted ...
written 11 weeks ago by Gavin Kelly180

Latest awards to Gavin Kelly

Scholar 11 weeks ago, created an answer that has been accepted. For A: DESeq2 3-fatcor design and different interaction terms
Teacher 11 weeks ago, created an answer with at least 3 up-votes. For A: extracting the genes associated with the clusters
Autobiographer 12 weeks ago, has more than 80 characters in the information field of the user's profile.
Scholar 18 months ago, created an answer that has been accepted. For A: DESeq2 3-fatcor design and different interaction terms

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 150 users visited in the last hour