This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Analyzing Data From Multiple Experiments

I've got microarray data from a bunch of patients. I wrote a Python script to compile the microarray/gene expression data into a manageable spreadsheet/table file so I could do some statistics in R. I also have a ton of other data (some from the same patient) including SNP's, methylation, copy numbers, etc. and was wondering if there was some standard method of analyzing all of this information simultaneously/together. I know I can analyze each set of data independently and then compare my results at the end, but I was wondering if there was something out there that does an integrated analysis. I found the Integrative Genomics Viewer which seems to help me visualize some of the data simultaneously, but I can't seem to find any analysis tools in this software that give me a quantitative difference between patients or groups of patients. My expression data is from a microarray and has already been mapped to genes, but some of the other data (SNP, methylation, etc.) only has position on a chromosome. I imagine that to compare my expression data with these other sets of data I would have to first map the chromosome positions to their respective genes. Any guidance that you could provide would be great since this is my first go at this type of problem. Thanks!

analysis genomics methylation snp expression
• 969 views • link
updated by seidel seidel • written by pstew pstew

This is not the type of question that we can help with. Combining data from different sources is a very complex task that critically depends on the particularities of your problem set.

• 0 views • link
updated by Obi Griffith Obi Griffith • written by Istvan Albert Istvan Albert

Ok thanks. I didn't think there was any sort of standard methodology for this but I wanted to ask to be sure.

• 0 views • link
written by pstew pstew

Hi Spamicles,

I dont think there is a standard solution to these kind of questions. The way different kinds of analyses should be intregrated depends on your research aims and questions. The Integrative Genomics Viewer provides a nice way for visualization of different kinds of genomic data, as well as Circos does. Using circos you can make eye-candy visualization that integrate the results of many kinds of genomic data.

See http://circos.ca

• 0 views • link
written by Irsan Irsan

I'll check this out. Thank you very much!

• 0 views • link
written by pstew pstew

1 answer

I don't know of any "standard method of analyzing all of this information simultaneously/together". You've got a complex data set, and you have many complex problems you could address with the data. I disagree and agree with Istvan's comment, in that we can help you with this question by pointing out that analyzing a data set such as yours is a very complex undertaking that depends critically "on the particularities of your problem set." If you know R, and you know how to do statistics in R, then I assume you know the caveats of dealing with microarray data from various sources/patients. As for the other data sets, mapping SNPs and other genomic events to genes or other loci is trivial in R using tools from bioconductor - so that should be straightforward for you to generate. Asking and answering questions about gene expression and genomic events from your data compilation....that's difficult without more specifics. You've got complex questions, there's no easy solution.

• 0 views • link
written by seidel seidel

I agree with you, the complexity level will depend on the experiment(s), and for sure you can use R:biocondutor to "integrate" the sources but isn't a trivial task.

• 0 views • link
written by JC JC

Thanks for the help. These replies were very helpful.

• 0 views • link
written by pstew pstew

Log in to answer this question.

More posts like this