Back

High-accuracy structure modeling for antibody-antigen complexes

Wang, S.; Zhuang, J.; Cui, X.; Lv, Z.; Hou, D.; Zhang, G.

2025-11-04 bioinformatics
10.1101/2025.11.03.686275 bioRxiv
Show abstract

Accurate modeling of antibody-antigen complexes is crucial for advancing therapeutics and diagnostics, yet predicting their binding interface remains a formidable challenge. To address this, we introduce DeepAAAssembly, a protocol that enhances static structural information from AlphaFold3 by integrating dynamic interaction patterns to guide complex assembly. Our approach leverages predicted inter-chain residue distances to construct a flexibility-aware energy function, which drives a two-stage conformational sampling process for global exploration and local exploitation. On a benchmark set of 67 representative antibody-antigen complexes, by incorporating a built-in confidence selection mechanism, DeepAAAssembly outperforms AlphaFold3, achieving not only a 12.9% higher average DockQ score and more medium- and high-quality models, but also reliably elevating the most challenging cases from incorrect to acceptable accuracy. These results demonstrate that DeepAAAssembly effectively captures conserved interaction motifs and conformational flexibility, offering a robust framework for high-accuracy antibody-antigen modeling.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

1
Nature Communications
○ 5641 papers in training set
Top 9%
18.4%
Similar papers in this journal
2
Journal of Chemical Information and Modeling
● 238 papers in training set
Top 0.4%
12.6%
Similar papers in this journal
3
Briefings in Bioinformatics
◐ 354 papers in training set
Top 1%
7.2%
Similar papers in this journal
4
Journal of Chemical Theory and Computation
● 140 papers in training set
Top 0.4%
5.5%
Similar papers in this journal
5
Communications Biology
○ 993 papers in training set
Top 2%
5.5%
Similar papers in this journal
6
Nature Machine Intelligence
○ 70 papers in training set
Top 0.4%
5.4%
50% of probability mass above
7
PLOS Computational Biology
● 1863 papers in training set
Top 9%
4.0%
8
Bioinformatics
◐ 1204 papers in training set
Top 5%
4.0%
9
Advanced Science
○ 286 papers in training set
Top 2%
4.0%
10
Computational and Structural Biotechnology Journal
● 242 papers in training set
Top 2%
3.2%
11
eLife
● 5828 papers in training set
Top 38%
3.1%
12
mAbs
○ 32 papers in training set
Top 0.2%
2.4%
13
Nature Methods
○ 385 papers in training set
Top 4%
2.1%
14
Protein Science
○ 246 papers in training set
Top 2%
1.7%
15
Scientific Reports
○ 3612 papers in training set
Top 52%
1.7%
16
Cell Systems
○ 201 papers in training set
Top 3%
1.7%
17
Structure
○ 193 papers in training set
Top 2%
1.3%
18
Proteins: Structure, Function, and Bioinformatics
○ 88 papers in training set
Top 1%
1.0%
19
Proceedings of the National Academy of Sciences
● 2444 papers in training set
Top 39%
0.9%
20
Journal of Molecular Biology
○ 232 papers in training set
Top 4%
0.8%
21
The Journal of Physical Chemistry B
● 167 papers in training set
Top 2%
0.8%
22
Cell Reports Methods
○ 165 papers in training set
Top 4%
0.8%
23
Bioinformatics Advances
◐ 203 papers in training set
Top 5%
0.6%
24
iScience
○ 1154 papers in training set
Top 41%
0.6%
25
Nucleic Acids Research
◐ 1281 papers in training set
Top 15%
0.6%
26
Communications Chemistry
○ 48 papers in training set
Top 2%
0.6%

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.