Unambiguous Assembly Of Next-Gen Fastq Reads Into Fastq Contigs?

Question

[2184687-1231-83-]

Hi,

Does anybody know of any tool that will produce an unambiguous assembly of next-gen fastq files and give the assembled output back as fastq with consensus/combined scores?

By unambiguous I mean something like this in abyss:

ABYSS -k$k -b0 -t0 -e0 -c0

Cheers,
Albert.

Answer 1

[Istvan Albert]

I think combining the quality score is not the intended use case for the p-values. The quality scores refer to the probability of a basecall being incorrect. So during assembly it is an important information.

But once the reads are assembled if we were to keep it, it would need to reflect not just the chance of one base being wrong but that for an entire overlap to occur by chance. That is a different quantity than the original.

This is just an opinion, I could be wrong.

Answer 2

[Peter]

Roche's "Newbler" gsAssembler and MIRA will both output FASTA+QUAL files for their contigs, which you can turn into a Sanger FASTQ file. They also output ACE files for the contigs which can also be converted into FASTQ files.

For MIRA the contigs have PHRED quality scores as high as 90, while for the Newbler assembly I just looked at the scores went up to 64. These scores are much higher than any single read (typically maximum PHRED 40) because they are based on a consensus.

You can probably do this for other quality aware assembler output too.

[2184687-1231-83-]

Great! Thanks! Any that will do it for Illumina GA2 reads?

[2184687-1231-83-]

Sorry, I just found out that MIRA claims to work for Illumina GA2 as well. Thanks!