Varscan Snp Strand Information

Question

[GPR]

Hello, I am moving through an SNPs analysis for the first time and am now looking at my VarScan outputs, with multiple questions. My RNA-seq data is paired-end with high coverage (>200 million reads) but not strand-specific. My goal is to compare the global A-to-G editing events in a pair of isogenic cell lines. After removing the known SNPs documented in dbSNP, I have extracted variant calls in Alu repeats, and thereafter focused on the A-to-G variants, leaving behind everything else.

My questions are: 1.I understand that VarScan reports in one strand only (first strand). Does this mean that T-to-C variant calls represent A-to-G SNPs in the reverse strand? 2.I have found editing sites in a group of genes in one sample, but not in the other one being compared. Given that this finding repeats in biological replicates, is this enough to conclude the absence of a given editing in one sample but not the other? 3.How much of RNA-editing sample variation should I expect? This is, compared to total mRNA or isoform levels by RNA-seq.

Thanks, G.

Answer 1

[Sean Davis]

Answer to question 1: Yes

Answer to question 2: I think what you are calling "editing events" can only be called putative editing events. You would need DNA to "prove" that these are, indeed, editing events. Be sure to read this article, though, as you are dealing with a truly hard problem:

http://nar.oxfordjournals.org/content/early/2013/01/08/nar.gks1443

[GPR]

Thanks Sean, your comments have been very useful. I have no come to realize that unlike SILAC proteomics or RNA-seq mRNA levels, RNA-editing my be a hard call in a quantitative comparative study.

Answer 2

[dankoboldt]

Thanks for the question, and to Sean Davis for providing an excellent answer!