Different k-mer combination generate different results

[Niohuruzh]

Description of bug

Hello,
I try to find a gene whether is complete in my genome assembled by SPADES based on Illumina data. First I try spades.py -k 21,33,55,77,99 --isolate -t 46 --cov-cutoff auto -1 ${a}_1.rd.fastq -2 ${a}_2.rd.fastq -o output to generate a genome. Then I use blastn, and I get a complete gene sequence. Second, I try spades.py -k 17,33,43,55,77,99,111 --isolate -t 46 --cov-cutoff auto -1 ${a}_1.rd.fastq -2 ${a}_2.rd.fastq -o output to generate a genome. Then I use blastn, and I get a fragmental gene sequences in different 'NODE'. So is there any suggestion to choose the best k-mer combination when assemble a genome.

Looking forward your reply!
Best wishes

spades.log

if need

params.txt

if need

SPAdes version

v3.15.4

Operating System

MINT20

Python Version

Python 3.7.12

Method of SPAdes installation

conda

No errors reported in spades.log

• Yes

[asl]

How long are your reads? And what is the average coverage? In general, we suggest to stick to defaults unless one is 100% sure what they are doing and what impact of the assembly is done

[asl]

Well, using such big k-mer length for such short reads is not recommended. Just use default k-mer lengths auto-selected by assembler

[Niohuruzh]

So which options --isolate and --careful are you recommended to use in this case?

[asl]

Start with --isolate

[Niohuruzh]

Thanks. But why not --careful? Because I have high-coverage data?

[Niohuruzh]

OK. It's not SPAdes problem. I use the wrong DNA sequences to blastn. Sorry to waste your time. Thanks so much!!!