Assembly and consensus tools for Nanopore reads from amplicons

Question

[Vitis]

I had Nanopore reads from amplicons as short as 600bp, and tried to do a quick and dirty assembly and consensus process. Both miniasm + racon and canu failed, suggesting that the reads were too short and there was no enough information for assembling. I wonder whether there is a tool that could do this. All needed is a quick overlapping process to put the reads together so we can tackle the random errors and generate relatively clean consensus sequences.

Answer 1

[Vitis]

Maybe it is valid to answer myself and this may help others in similar situations. Tweaking some canu parameters would do the job:

canu -p candidate -d candidate useGrid=0 -nanopore-raw candidate.fa genomeSize=1000 stopOnReadQuality=false minReadLength=50 minOverlapLength=20 corMinCoverage=3

Minimum read length, minimum overlap and minimum coverage could all be tuned to fit different case. Sometimes it stops at the correction step, the longest correct read would cover all other reads so essentially there was no assembly to be done. The hints came from canu GitHub site.

https://github.com/marbl/canu/issues/697