Question: minfi sample limit
0
3.7 years ago by
igor30
United States
igor30 wrote:

I recently started using minfi for analyzing a large set of 450k methylation arrays. For a small trial run, everything worked as expected, but once you get over around 200 samples, there are some problems. Some functions like getBeta() work as expected. However, qcReport() and densityPlot() skip some samples. I don't get any errors. With density plots, it's hard to notice that any are missing, but with the bean plots, certain samples display the sample name, but the graph is blank. Changing the group labels changes which samples show up, so it's not a problem with the samples or the input object. Has anyone else run into this problem? Now I am concerned that other functions may be affected, but I am just not noticing it.

modified 3.7 years ago by Kasper Daniel Hansen6.4k • written 3.7 years ago by igor30
0
3.7 years ago by
United States
Kasper Daniel Hansen6.4k wrote:
This sounds really weird. I suspect your sample sheet is messed up. Kasper On Thu, Mar 17, 2016 at 10:59 AM, igor [bioc] <noreply@bioconductor.org> wrote: > Activity on a post you are following on support.bioconductor.org > > User igor <https: support.bioconductor.org="" u="" 7184=""/> wrote Question: > minfi sample limit <https: support.bioconductor.org="" p="" 79762=""/>: > > I recently started using minfi for analyzing a large set of 450k > methylation arrays. For a small trial run, everything worked as expected, > but once you get over around 200 samples, there are some problems. Some > functions like getBeta() work as expected. However, qcReport() and > densityPlot() skip some samples. I don't get any errors. With density > plots, it's hard to notice that any are missing, but with the bean plots, > certain samples display the sample name, but the graph is blank. Changing > the group labels changes which samples show up, so it's not a problem with > the samples or the input object. Has anyone else run into this problem? Now > I am concerned that other functions may be affected, but I am just not > noticing it. > > ------------------------------ > > Post tags: minfi, illumina, illimina 450k methylation, methylation > > You may reply via email or visit minfi sample limit >

If the sample sheet is messed up, then subsetting it to problematic samples should produce problems again. However, if I subset the targets data frame to 20 samples that got skipped in a set of 200 and just use that, everything is fine.

It is not clear that what you say is true. Do you have any duplicates in the columns identifying the individual samples, for example the Basename column? On Thu, Mar 17, 2016 at 11:33 AM, igor [bioc] <noreply@bioconductor.org> wrote: > Activity on a post you are following on support.bioconductor.org > > User igor <https: support.bioconductor.org="" u="" 7184=""/> wrote Comment: minfi > sample limit <https: support.bioconductor.org="" p="" 79762="" #79764="">: > > If the sample sheet is messed up, then subsetting it to problematic > samples should produce problems again. However, if I subset the targets > data frame to 20 samples that got skipped in a set of 200 and just use > that, everything is fine. > > ------------------------------ > > Post tags: minfi, illumina, illimina 450k methylation, methylation > > You may reply via email or visit > C: minfi sample limit >