I have been visualising number of alignments to each nucleotide using a depth file generated from minimap2 alignment to a reference then bedtools genomecov:
bedtools genomecov -strand + -ibam REF.sorted.bam -g REF -d > REF.d
I am finding that there are occassional drops in depth across the genome: 20-80% decrease in depth at several regions of 1-3 nucleotides.
I am wondering a few things:
- what could be the cause of these drops?... systematic sequencing errors (it is nanopore sequencing data, not great at homopolymer sites), could it be an alignment artifact, or error in reference sequence.
& more pressingly:
- when I visualise this using tablet, those drops are not in the visualisation - is it common to "scrub" these drops out of depth visualisations?
If so, can anyone point me to a paper or github code that describes a reliable approach to take... Would like to find out since at the moment the depth histograms I generate look very bumpy. Are there tools out there for generating "scrubbed" depth visualisations.