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).
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)
In my opinion, RNA-seq can be a complementary method to extend DNA variant analysis:
You can observe how much a variant is expressed due to different allelic gene expression.
By looking at coverage levels, you can detect exon deletions due to intronic splicing variants. With WGS you are able to detect those splicing variants, but frequently you are not sure if that variant is going to cause a splicing event or not.
You can discover alternate/new transcripts that correspond to different proteins.