Align concatameric reads to region

Question

[conradinbaumi]

Hi guys,

I want to ask if there is anything out there that might help me do this before I start to spend too much time trying to hack it together. I have a set of very long >10kb nanopore reads that theoretically contain the same 4kb region within over and over in the same orientation. Is there an easy way to align all of the individual repeats of this region to said region?

I was thinking of iterating through all reads, do a local alignment of the region and store the aligned part in a sub read set, hence computationally fragmenting all of the reads. But maybe there is a more elegant way of doing this?

Thanks for all the advice Conradin

Answer 1

[stolarek.ir]

You can google under nanopore circular read. Rolling circle amplification is used in nanopore to mimic the pacbio sequencing process.

In this publication they processed the data which resemble your problem:

https://www.pnas.org/doi/10.1073/pnas.1806447115#sec-3

""""" The following steps are performed to generate a consensus sequence: (i) Tandem repeats in each raw read are detected using a modified EMBOSS WATER (21–23) Smith–Waterman self-to-self alignment. (ii) Raw reads are then split into complete subreads containing full repeats and incomplete subreads containing partial repeats at the read ends. If there is more than one complete subread, a preliminary consensus of these complete subreads is computed using poaV2 (15). If only one complete subread is present in the raw read, its sequence is used as consensus in the following steps. (iii) Complete and incomplete subreads are aligned to the preliminary consensus sequence using minimap2 (24). (iv) These alignments are used as input to the racon (16) algorithm, which error-corrects the consensus. """"