Dear All,
What is the advantage of variant calling from transcriptomic sequence (RNA-seq) compared to whole genome sequencing? Whether we will lose any valuable information while using RNA seq for variant calling?
The only benefit to variant calling from RNAseq is that you can focus on genes/transcripts that are actually expressed. In all other respects doing this is at best equivalent and typically worse than standard variant calling on WGS data. I would strongly encourage you to perform variant calling on DNA and then filter results for relevance based on gene expression in your tissue/system of interest. You then don't run the risk of losing variants due to mono-allelic expression or any of the many other biases one can think of when using RNA-seq for variant calling (see the links on the right-hand side of the page for further discussion).
Yes, though (A) you will likely not have coverage of all of the GWAS variants if you use RNAseq data and (B) RNAseq variants will be more likely to be called homozygous or have other odd quality properties due to extreme expression changes. So take any comparisons with an appropriately sized grain of salt.
There is no advantage of RNA-seq variants over WGS, because RNA-seq is the more complicated protocol due to the reverse transcription (writing RNA into cDNA prior to preparing the actual sequencing library, which can introduce PCR-mediated base pair errors). How can you validate, well that depends if you downloaded the data or if you have the original RNA/DNA to do some Sanger sequencing. If you downloaded, you cannot.
Much like anything in Genomics and Bioinformatics, it really depends on what your question is. While there's no technical advantage in choosing RNAseq variant calling compared to WGS to identify disease-causing variants, RNAseq is an order of magnitude cheaper. It could be an option if you are looking at large populations or running some sort of eQTL or GWAS analysis. There are workflows which achieve pretty good results when applying variant imputation with RNAseq or low-coverage genomic sequencing (or RAD-seq/GBS)
While there's no technical advantage in choosing RNAseq variant calling compared to WGS
I disagree. RNAseq has a number of disadvantages, particularly if you're gene of interest isn't heavily expressed in the tissue you have to assay (e.g., it's tough to get brain tissue from people, but really easy to get a blood sample). Power in variant calling increases with sequencing depth, which can then be compromised. The additional random biases one also encounters in RNAseq ("random hexamer priming", 3' bias due to poly-A enrichment, RNA degradation, etc.) further complicate things.
WGS is more expensive, but it's the right tool for the job.
In my opinion, RNA-seq can be a complementary method to extend DNA variant analysis:
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
Sharing a paper
Thanks for the paper. I am working on alzheimers disease transcriptome data. After variant calling i got 480 variants using multiple filtering criteria. I compared the predicted variants with various GWAS and eQTL studies (AD,PD,ALS,AMD). I got some 40 variants are novel which are not reported in these studies. most of the predicted variants from intronic region. How i will validate this novel variation? Is it really novel variants or artifacts Note: sample size control:5, AD:5 alignment rate :~90 Quality considered for variant calling: Q-30 FDR<0.05 Read depth:minimum-10
Please help me.