Lost plasmids

[valery-shap]

Hi,

Thank you for developing and maintaining this very useful tool.

I am testing Hybracter on a Klebsiella pneumoniae isolate that contains several relatively large plasmids, and I observed a discrepancy compared to other assemblies.

Background

Using Flye (standard mode) and Dragonflye, I consistently recover two plasmid groups:

IncFIB(Mar)_1_pNDM-Mar and IncHI1B_1_pNDM-MAR
IncFIB(K)_1_Kpn3, IncFII_1_pKP91, and IncR_1

Flye assembly summary:
#seq_name length cov. circ. repeat mult. alt_group graph_path
contig_1 5395890 49 Y N 1 * 1
contig_3 274838 45 Y N 1 * 3
contig_2 231095 15 Y N 1 * 2

However, when running Flye with --meta, an additional important plasmid (IncL/M) is recovered (carrying ARGs, although split across contigs). I suspect this plasmid is missed in standard mode due to low coverage:

#seq_name length cov. circ. repeat mult. alt_group graph_path
contig_2 5396077 57 Y N 1 * 2
contig_5 274994 53 Y N 1 * 5
contig_1 231139 19 Y N 1 * 1
contig_4 61345 9 N N 1 * *,4
contig_3 3277 11 N N 1 * 3

Hybracter results

I then ran Hybracter v0.12.0 in long-single mode.

Only one plasmid group was fully recovered:

IncFIB(Mar)_1_pNDM-Mar and IncHI1B_1_pNDM-MAR

Output (sample_per_contig_stats.tsv):

contig_name contig_type length gc circular
chromosome00001 chromosome 5396075 57.4 True
plasmid00001 plasmid 274991 46.65 True
plasmid00002 plasmid 37652 52.12 False
plasmid00003 plasmid 2623 56.04 True

Question

Could you clarify what might explain this behavior?

Thank you very much for your help!

Best regards,
Valery