Any recommended parameter-set to adjust the low coverage hifi data(15~20X)

[chhylp123]

Any recommended parameter-set to adjust the low coverage hifi data(15~20X)

[jelber2]

You might have to simulate HiFi coverage. For example get a similar haploid genome assembly to what you have not yet assembled, then mutate it into diploid (or whatever your ploidy is) with mutate.sh from BBTools/BBMap (https://sourceforge.net/projects/bbmap/files/) with some predicted "heterozygosity rate". Then simulate random reads with randomreads.sh (also from BBTools/BBMap) using PacBio HiFi characteristics (12-15Kbp, error rate 1.0-1.5% or whatever you desire) such as the example commands below. Note that I put coverage=10 because there are actually two haplotypes produced by mutate.sh when ploidy=2, so that 10x diploid coverage equals 20x haploid coverage.

mutate.sh in=curated.fasta ploidy=2 maxindel=5 subrate=0.0002 hetrate=0.05 indelrate=0.00002 out=curated.fasta.diploid.mut.fasta vcf=curated.fasta.diploid.mut.vcf.gz 2>mutate.log &
randomreads.sh ref=curated.fasta illuminanames=t coverage=10 gaussianlength=t pacbio=t pbmin=0.01 pbmax=0.015 out=random_reads_ccs.fasta minlength=10000 midlength=12000 maxlength=15000 > random_reads_ccs.log 2>&1

Then use QUAST-LG (https://github.com/ablab/quast) to compare you hifiasm results on different values of k,r, and a with curated.fasta.diploid.mut.fasta if you use unitigs or curated.fasta if you use primary contigs (just a suggestion).

[chhylp123]

I agree with @jelber2. Thank you so much.

[KaiqiangLiu]

Thanks for your @jelber2 useful suggestion with all the details.
Thanks for your @chhylp123 reply.