Mabs, a suite of tools for gene-informed genome assembly

Abstract

Motivation: Despite constantly improving methods of genome sequencing, the task of error-free eukaryotic genome assembly is not yet solved. Among other kinds of problems of eukaryotic genome assembly are so-called "haplotypic duplications", which may manifest themselves as cases when alleles are mistakenly assembled as paralogs. Haplotypic duplications are dangerous as they create illusions of gene family expansions and, thus, may bring a scientist to wrong conclusions about genome evolution and functioning. Results: Here, I present Mabs, a suite of tools that serve as parameter optimizers of popular genome assemblers Hifiasm and Flye. By optimizing parameters of Hifiasm and Flye, Mabs tries to create assemblies where genes are assembled as accurately as possible. Tests on 6 eukaryotic genomes showed that in 6 out of 6 cases Mabs created assemblies with more accurately assembled genes than Hifiasm and Flye did when they were run with default parameters. When assemblies of Mabs, Hifiasm and Flye were postprocessed by a popular tool for haplotypic duplications removal Purge_dups, genes were better assembled by Mabs in 5 out of 6 cases. Availability and implementation: Mabs has been written in Python and is available at https://github.com/shelkmike/Mabs.

The genome of the toxic invasive speciesHeracleum sosnowskyicarries an increased number of genes despite absence of recent whole-genome duplications

“…To overcome this problem, we used the tool Mabs. Mabs is an add-on to long-read assemblers that allows automatic optimization of assembly parameters (Schelkunov, 2022). The application of Mabs improved the assembly both in terms of N50 and a fraction of duplicate genes (Table 1, Supplementary Table A2).…”

“…Oxford Nanopore reads were trimmed by Porechop (Wick, n.d.). The genome of H. sosnowskyi was assembled by Mabs-hifiasm (Schelkunov, 2022), Hifiasm (Cheng et al ., 2021), Flye (Kolmogorov et al ., 2019), Miniasm (Li, 2016), and Canu (Nurk et al ., 2020). After assembly, haplotypic duplication removal was performed by Purge_dups (Guan et al ., 2020).…”

“…After assembly, haplotypic duplication removal was performed by Purge_dups (Guan et al ., 2020). To compare the gene assembly quality in these assemblies, BUSCO (Simão et al ., 2015) and calculate_AG (Schelkunov, 2022) were used. The assembly with the most accurately assembled genes was the assembly made by Mabs without Purge_dups; this assembly was used in further analyses.…”

The genome of the toxic invasive species Heracleum sosnowskyi carries an increased number of genes despite absence of recent whole‐genome duplications

“…We found that Hifiasm performed better than other assemblers; however, it produced many duplicated genes due to incomplete collapsing of haplotypes. To overcome this problem, we used the tool Mabs—an add‐on to long‐read assemblers that allows automatic optimization of assembly parameters (Schelkunov, 2023). The application of Mabs improved the assembly both in terms of N50 and a fraction of duplicate genes (Table 1, Table S2).…”

“…Oxford Nanopore reads were trimmed by Porechop (Wick, n.d.). The genome of H. sosnowskyi was assembled by Mabs‐hifiasm (Schelkunov, 2023), Hifiasm (Cheng et al., 2021), Flye (Kolmogorov et al., 2019), Miniasm (Li, 2016), and Canu (Nurk et al., 2020). After assembly, haplotypic duplication removal was performed by Purge_dups (Guan et al., 2020).…”