Indeed, the IGV screenshot is fine. This is why I was asking how many reads as % are outside of exons. Actually, the only way to confidently know that a read is RNA is if it spans an exon-exon junction or the paired-end mate clearly jumps an intron due to splicing. Pure exonic or intronic reads can come from DNA as well, so it is really a matter if you have reads from outside annotated exons and gene bodies.
RNA-Seq with DNA contamination, any way to salvage the data?
Hello,
My group has performed RNA-Seq using Illumina Truseq RNA Exome kit, we performed alignments with HISAT2 and we noticed that a good portion of our samples had DNA contamination. So far, I have seen a lot of discussion on identifying the presence of DNA contamination in RNA-Seq data, but not a lot of discussion on what to do afterwards.
What we are most concerned about are contaminating DNA reads that align to the exome regions. Is there a way to salvage these samples by removing the contaminated DNA reads?
Below is an image from IGV showing the alignments of two samples.
Looking forward to any suggestions for this!
1 answer
Depending on your goal... you could certainly align to the transcriptome and discard the unaligned reads. Then realign the remainder to the genome. Of course, that assumes you're not trying to discover new exon junctions.
In that image I don't see any evidence of DNA contamination, though. If there was, I'd expect to see reads mapping at random locations throughout the intron, which is not happening. Instead, I'm seeing lots of PCR duplicates and reads that align into the introns with indels, and possibly the typical rainbow of SNPs indicative of a bad alignment, but are not present in the image because presumably your mismatches are not being colored in the reads. Maybe this is a situation where usually one splice site is used, but occasionally another one a little farther out is used? Or maybe the reads are just being misaligned for some reason and they actually do span the intron? Turning on SNP coloring might be handy.
• 0 views
•
link
written
by
Brian Bushnell
Log in to answer this question.
More posts like this
-
High percentage of intronic/intergenic reads in RNA-seq
written by lluc.cabus •
-
bulk RNA-seq contamination, lung ambient RNA
written by Lpak •
-
Mutect2 "somatic" variants with resistant tissue culture lines
written by shawn.w.foley
-
Genome-aligned RNA-Seq: one treatment showing high% of unmapped reads:too short
written by Biogeek
-
Where to download Normal exome bam for creating PoN (Panel of Normals) for cancer CNV workflow
written by prabhakar8279 •
-
low mappability in single cell rna-seq data?
written by ccagg
-
How to remove contamination from NGS data
written by olp123 •
-
Tool: Human NGS Cancer Data for tool development, algorithm benchmarking, teaching, pipeline evalua…
written by Malachi Griffith
-
Illumina Stranded Rna-Seq Background Problem
written by Nicolas Rosewick
-
What Could Be The Reason For Spliced Alignments In Chip-Seq Data?
written by Mikael Huss
How many reads in total (percentage) map outside of annotated exons?