Identifying a true viral genomic deletion, how to distinguish real deletions from quasispecies noise in IGV coverage gaps
As a beginner in viral genomics, I map raw reads to a reference genome, extract virus-specific reads, filter for quality, and perform de novo assembly, resulting in two contigs separated by a gap of several hundred base pairs — a pattern consistent across all samples. In IGV, this region shows a sharp drop in coverage, suggesting a deletion, yet a small number of reads still map across the gap, raising the question of whether this reflects a true genomic deletion or residual quasispecies variation.
My questions are:
- Is it correct to interpret this pattern — a sharp drop in coverage with residual mapping reads — as a true genomic deletion? This pattern is reproducible across all my samples.
- Are the reads that still map across this region likely to be sequencing/alignment noise, or could they represent biologically relevant quasispecies (i.e., variants retaining the sequence)?
- If this is a deletion, what is the most accurate computational method to define its precise 5’ and 3’ boundaries without resorting to PCR or Sanger sequencing?
0 answers
No answers yet.
Log in to answer this question.
More posts like this
-
Read Mapping of Reads to Reference Genome Shows Extreme Bias
written by Lawrence •
-
High amount of intronic/intergenic reads in SMARTer stranded total bulk RNAseq
written by Mat •
-
Need help with IGV interpretation
written by Hyper_Odin
-
Extracting internal reads from SmartSeq3 data using zUMIs
written by pm2012
-
Whole intron deletion?
written by German.M.Demidov
-
Strange coverage of mtDNA - need help interpreting MiSeq results
written by tatiatcg •
-
discrepancy between reads
written by n.mekkes •
-
Higher coverage when using genomic sequencing reads from a pool of individuals
written by jomaco
-
Verifying/Investigating Frameshifts On A De Novo Assembly
written by Tancata •
-
Genomic Read Mapping Biased Towards Coding Regions?
written by Vitis
Is this a sample where more than one genome is expected (e.g. host + virus). If so are you aligning to both genomes at the same time?
If this is a single genome sample, are you aligning to a related/close genome?
Yes, both virus and host are present, but I first map all reads to the viral genome and select only those that successfully map. The screenshot in IGV shows these selected reads remapped to the reference. The reference is very closely related—nearly identical in sequence to the virus in my sample.
In theory if both genomes are present in the sample then you should map to both at the same time. There is some chance that reads from host that have sequence similarity to the virus could get mapped (aligners make a best effort to map reads even though they may have originated from a different genomes). You could also use a tool like BBSplit ( https://bbmap.org/tools/bbsplit ) that allows one to make intelligent decisions about reads that map to multiple genomes.
Considering the difference in coverage it is possible that the deletion is real and the reads you see aligning there is noise. Computational methods are not going to give an absolute answer. Confirmatory sequencing with sanger/PCR may be the conclusive proof needed to delineate the ends of the deletion.