Different result in same species in jellyfish(genome size estimation)

Question

[ylkim9]

Hello, I have some questions about using jellyfish.

Recently I estimated genome size with jellyfish with 2 species (17 samples in A species and 2 samples in B species). I set k-mer with 19 and 25. The reason i set 25mer is that my labmate had a normal result(that we know generally) with other species in same genus. But I had only normal result with 2 samples in 19mer, 2 samples 25mer in A species and only one sample in 19mer in B species.

Do I have to set another k-mer or do it again with same sets?

[dthorbur]

Please provide more context. How far out of your expected range were the results? This could simply be a case of a priori assumptions being wrong.

Regardless, repeating analyses with different parameters is usually an acceptable way to try and see if a putative erroneous output is related to a specific parameter setting.

[ylkim9]

The genome size of A, B species are both 1.0 ~ 1.1 gbp. However, wrong results showed me about 0.1~0.4 gbp and 60~85% model fit. I also tried 17mer and 31mer but all wrong result showed as same as 19 and 25mer's wrong results.

I used command like this.

NCPUS=70
 jellyfish count -t $NCPUS -C -m 19 -s 5G -o $OUTDIR/${ID}.19mer_out --min-qual-char=? <(zcat $FQ1) <(zcat $FQ2)
    jellyfish histo -t $NCPUS $OUTDIR/${ID}.19mer_out > $OUTDIR/${ID}_19mer.histo
    genomescope2 -i $OUTDIR/${ID}_19mer.histo -o $OUTDIR -k 19 -p 2

and the one of the wrong result summary file showed like this.

GenomeScope version 2.0 input file = path/to/inputfile/CNGR_1.gz_17mer.histo output directory = path/to/outputfile/CNGR_1.gz_17_ploidy p = 2 k = 17

property min max
Homozygous (aa) 89.553% 92.5941%
Heterozygous (ab) 7.40594% 10.447%
Genome Haploid Length 406,470,818 bp 416,867,094 bp
Genome Repeat Length 290,706,938 bp 298,142,329 bp
Genome Unique Length 115,763,880 bp 118,724,765 bp
Model Fit 36.6236% 84.2415%
Read Error Rate 0.756208% 0.756208%