Nanopolish without reference genome

[jordi]

Hi all!
I am trying to correct raw MinION sequencing reads with nanopolish since nanocorrect seems to be deprecated.
However, during the nanopolish workflow found here an alignment to a reference genome seem to be compulsory. So, how to proceed with nanopolish lacking a reference genome? I mean, how to correct raw sequencing reads (or assembled reads) without the bam file obtained with bwa?
Any help will be really apreciated.

[cstack]

Hi all!
I am trying to correct raw MinION sequencing reads with nanopolish since nanocorrect seems to be deprecated.
However, during the nanopolish workflow found here an alignment to a reference genome seem to be compulsory. So, how to proceed with nanopolish lacking a reference genome? I mean, how to correct raw sequencing reads (or assembled reads) without the bam file obtained with bwa?
Any help will be really apreciated.

I guess you could try an all-vs-all mapping (skipping self-mapping) -- I think that graphmap and minialign can do this and output the results in SAM/BAM format.

[apredeus]

There is a tool from Schatz lab called NanoCORR: http://schatzlab.cshl.edu/data/nanocorr/

I don't think it's used much though. Also, as far as I know, nanopolish is the only tool that uses fast5 information. Do you need just the nanopore reads, not the assembly?

I would then suggest making an assembly, correcting it with nanopolish, then aligning the reads to the assembly and extracting the result from the BAM file. Those won't really be "reads" but they will be read-like fragments of the polished genome, which is roughly equivalent to corrected reads.

Otherwise, if you also got the Illumina data, you can use Masurca and find the "mega-reads" it generates. Those are also long reads corrected using Illumina before the assembly.