Entering edit mode
4.4 years ago
rprog008
▴
70
Dear member,
I am new to copy number variations analysis and thus seeking help how to start copy number variation analysis in r using the table containing column data as given below.
sampleid Chr POSITION_START POSITION_END REF ALT refGene CNV_log2
I shall be highly obliged if people working in the same field suggests me some good papers which i may follow including packages in R which might help me.
Thanks in advance
As I understood, you already have CNVs called in your dataset. Then there are thousand of ways how to analyse this data. The question that we need to know: which scientific question you try to answer?
Yes I have cnv dataset and I am trying to find a general pattern in my CNV dataset, for instance, performing PCA, clustering and see which cnv is related with each other and so on. But, as I stated earlier, i am new to this field and thus seeking advise from expertise, like you, what are the important scientific questions i can address with my dataset
That does not seem like a clear scientific question or hypothesis (?) Were you simply given this dataset and told to process it (in any way)? - no guidance from any supervisor?
If this is purely about learning how to do clustering, PCA, etc., then, as opposed to one of us providing the entire solution in this single thread, I suggest that you spend a week or so researching / searching how to do these - there are many tutorials and postings about these topics all across the World Wide Web.
Yes. at present I am not associated with any institute or supervisor. Just looking for a suitable job/phd after my master. As my master supervisor is not that much expertise in this field and thus I am looking for www everywhere since last one month but unable to get a proper link and thus asking member to help me with some important tutorial and paper which I may follow.
IMO copy-number data is not something super different from diff expression data or SNVs data in terms of visualization - so simply take any manual from the online and do same things as for other types of data. The simplest important scientific question you may rise is "do I have CNVs that are associated with some phenotype in my cohort". The simplest phenotype is ancestry - check if you can separate your samples based on their origins.
I see, but, do you specifically need to use CNV data or can you just view it as any data matrix of numbers?
You could look through the Bioconductor Workflows:
Thanks @ Kevin Blighe and @ German.M.Demidov. Your help really means a lot to me. I am grateful for your kind act.
@ Kevin Blighe, Yes, I am working with CNV data because professor under whom I am interested to persue my phd works with CNV data. After learning the analysis I will write to him.
The important question is are your CNVs germline or somatic. If germline - you may want to compare them with https://www.ncbi.nlm.nih.gov/pubmed/26432246 . If somatic - with TCGA pancancer atlas (e.g. marker paper - https://www.ncbi.nlm.nih.gov/pubmed/29625048 )
Mine CNVs are somatics in nature
There are various ways how to analyse somatic CNAs. Usually it is performed in a per-gene manner. Summarize your CNA data in a manner: gene - copy number for each sample (since they are long), make clustering, make PCA/MDS, try to correlate them with phenotype, try GISTIC2.0 analysis, do oncoplot, analyse cellularity / clonal structure of your tumors, study events such as firestorms, chromotripsis, chromoplexy, it should be enough.
@ German.M.Demidov.. Thanks a lot
I have also found one link and that's seems very interesting. I would like to share with you. https://bioconductor.org/packages/devel/bioc/vignettes/CNVRanger/inst/doc/CNVRanger.html
I am really thankful to you, Otherwise i was lost since one month
Hey @rprog008 I have never used this package but it does not look bad =) you can start with it