Back

Cross-architecture ensembling of DNA foundation models improves the precision and stability of chimera detection in long-read metagenomic bins

MinSeo, K.; Jae-Ho, S.

2026-07-07 bioinformatics
10.64898/2026.07.02.735979 bioRxiv
Show abstract

Motivation: Chimeric metagenome-assembled genomes (MAGs) that pool DNA from multiple organisms contaminate downstream analyses. Marker-gene tools such as CheckM2 miss low-level chimerism, and DNA foundation models have been proposed as a sequence-composition alternative, but whether large autoregressive models (Evo2, 7B parameters) outperform smaller contrastive models (DNABERT-S, 117M) has not been rigorously tested.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

1
Genome Biology
○ 637 papers in training set
Top 0.4%
16.7%
2
Nature Communications
○ 5641 papers in training set
Top 12%
14.8%
3
Nature Methods
○ 385 papers in training set
Top 0.9%
10.4%
4
Nature Biotechnology
○ 172 papers in training set
Top 0.2%
9.5%
50% of probability mass above
5
Bioinformatics
◐ 1204 papers in training set
Top 3%
7.7%
6
Briefings in Bioinformatics
◐ 354 papers in training set
Top 3%
3.4%
7
NAR Genomics and Bioinformatics
◐ 242 papers in training set
Top 2%
3.1%
8
PLOS Computational Biology
● 1863 papers in training set
Top 13%
2.1%
9
GigaScience
◐ 212 papers in training set
Top 2%
2.1%
10
Microbiome
○ 154 papers in training set
Top 1%
1.9%
11
Nature Computational Science
○ 55 papers in training set
Top 0.6%
1.7%
12
Nucleic Acids Research
◐ 1281 papers in training set
Top 9%
1.7%
13
Nature Genetics
○ 286 papers in training set
Top 3%
1.7%
14
Cell Systems
○ 201 papers in training set
Top 3%
1.5%
15
Genome Research
● 468 papers in training set
Top 4%
1.4%
16
Molecular Biology and Evolution
◐ 542 papers in training set
Top 4%
1.1%
17
Scientific Reports
○ 3612 papers in training set
Top 67%
1.1%
18
BMC Bioinformatics
○ 457 papers in training set
Top 5%
1.1%
19
Microbial Genomics
● 225 papers in training set
Top 2%
1.1%
20
Cell Reports Methods
○ 165 papers in training set
Top 3%
1.0%
21
PLOS ONE
● 5266 papers in training set
Top 58%
1.0%
22
BMC Genomics
○ 406 papers in training set
Top 7%
1.0%
23
Cell
○ 431 papers in training set
Top 10%
0.8%
24
Nature
○ 645 papers in training set
Top 10%
0.8%
25
Computational and Structural Biotechnology Journal
● 242 papers in training set
Top 8%
0.6%