Spades - min-contig threshold

[sav0016]

Hello,

I'm trying novel-x and I'm getting an error with Quast and Spades. Should I change the min contig threshold or kmer option in assembler? Quast and Spades logs attached.

quast.log
spades.log

Thanks!

Jakub

[1dayac]

Hello Jakub!
I feel that it is actually normal, some of the local assemblies don't have contigs longer than 500bp, e.g. they are spurious. Snakemake settings are set in the way that the error for a single local assembly doesn't generate an error for a whole run. Does it work for you?

Thanks,
Dmitry

[sav0016]

Hi,

thank you very much for quick reply!

I´m attaching snakemake log with error and also last 1000 rows of log file for run.

Novel-X-last1000.log
2020-11-29T180755.809346.snakemake.log

It won´t finish run.

Jakub

[1dayac]

At this point I wonder, what is the coverage of the bam file?
Novel-X works well when the coverage is around 50-60X (usually bam-file is >150Gb).
Such errors might occur if coverage is very low and we don't assemble any contigs >500bp at all

[sav0016]

I think coverage is around 30X ... Bam file is 100 GB. So is there any way how to analyze it? Thanks a lot!

Jakub

[1dayac]

Well... From my experience, NUI-pipeline still works well with this coverage - https://www.nature.com/articles/s41467-018-05513-w
The problem with NUI pipeline - it is not a real software, so you will need to replace some hard-coded paths, make code tweaks here and there. If you encounter problems that don't have obvious solutions, you can write me and I will try to help

[sav0016]

So what you recommend? Decrease contig filter length? Thanks.

[1dayac]

Unfortunately, Novel-X will not work on 30X coverage. Try NUI-pipeline, if you will be able to get results - they should be decent.