We have made a targeted assembly tool, that concentrates assembly based on targeted DNA sequences. ARGextender, available here: https://github.com/genomicepidemiology/ARGprofiler. Where KMA is used to identify the reads / read pairs matching the targets, which is then assembled with SPAdes. This process is then repeated iteratively till convergence, or a max number of iterations has been made. This did not exactly make the assemblies better, but they were produced much faster and with much lower memory peaks.

I am not really sure how they simulated the reads in this study, the method section is very sparse. It seems that several plasmids (with the same gene) were simulated, with what seems to be an insert length shorter than the amr genes. Hence, they are bound to break right at the edges of the amr genes and can't really do anything else. Some bridging can in theory be done based on depth, codon usage etc., but this it seems that only depth was used here (internally implemented in the assembly methods), upon which it is unsure whether the depth varied between the simulated plasmids within each sample.

[http://www.dtu.dk/-/media/DTU_Generelt/Andet/mail-signature-logo.png] Philip Thomas Lanken Conradsen Clausen Postdoc National Food Institute plan@food.dtu.dkmailto:plan@food.dtu.dk Kemitorvet Building 204 2800 Kgs. Lyngby www.food.dtu.dkhttp://www.food.dtu.dk/ ________________________________ Fra: Marco van Zwetselaar via Bioinfo List bioinfo-list@seqshare.org Sendt: 27. oktober 2024 00:29 Til: bioinfo-list@seqshare.org bioinfo-list@seqshare.org Emne: [Bioinfo-list] Metagenomic assemblies tend to break around antibiotic resistance genes

Abramova A, Karkman A, Bengtsson-Palme J. Metagenomic assemblies tend to break around antibiotic resistance genes. BMC Genomics. 2024;25: 959. doi:10.1186/s12864-024-10876-0https://doi.org/10.1186/s12864-024-10876-0

@Philip I seem to recall a paper from DTU on a tool that tried to solve precisely the issue your eastern neighbours stumble over in this paper?

Cheers Marco