alexvpickering/crossmeta

mixing our own .cel with GSE references

abasseville opened this issue · 3 comments

abasseville commented

Hi,
Thank you for your package, it is very usefull, and seams easy to use with GEO references.
I have a question (more than an issue).
I would like to mix some GSE references with my own .CEL and sample annotation.
I found your very nice website https://rnama.com/app/datasets, but I am only interested in downloading, normalizing, and "gene annotating" the mixed data (geo and mine). My purpose is to obtain at the end a single matrix of all samples after normalization (I am not interested by DE genes)
Is there a way to do that?
Thank you for your help

agnes

alexvpickering commented

Hi Agnes,
That would take some customisation to accomplish. I would approach this as follows:

  1. Use load_raw to load and annotate the raw GSE data. Dig into the source code of crossmeta::load_raw and adapt in order to do the same for your personal CEL files.

  2. Then change the row names to something that is common between all studies (e.g. the SYMBOL column in fData of each eset).

  3. Then merge all the matrices on those common identifiers.

abasseville commented

Hi Alex,
Thank you for your response.
I looked into the source code , but I am not sure I understood everything (I am rather beginner in bioinformatics): I couldn't find the line in load_raw function and in the linked-functions (load_affy, ...) where the "cross-platform effect size" is performed after the esets were properly loaded.
It seems for me that this step (which is es_meta if I understood well?) is done after the differential expression calculation .
Did I miss something?

abasseville commented

Sorry,I understood my mistake, I wanted to do cross-platform normalization and not meta-analysis.