Back

Generation of antigen-specific paired chain antibody sequences using large language models

Wasdin, P. T.; Johnson, N. V.; Janke, A. K.; Held, S.; Marinov, T. M.; Jordaan, G.; Vandenabeele, L.; Pantouli, F.; Gillespie, R. A.; Vukovich, M. J.; Holt, C. M.; Kim, J.; Hansman, G.; Logue, J.; Chu, H. Y.; Andrews, S. F.; Kanekiyo, M.; Sautto, G. A.; Ross, T. M.; Sheward, D. J.; McLellan, J. S.; Abu-Shmais, A. A.; Georgiev, I. S.

2025-02-17 bioinformatics
10.1101/2024.12.20.629482 bioRxiv
Show abstract

The traditional process of antibody discovery is limited by inefficiency, high costs, and low success rates. Recent approaches employing artificial intelligence (AI) have been developed to optimize existing antibodies and generate antibody sequences in a target-agnostic manner. In this work, we present MAGE (Monoclonal Antibody GEnerator), a sequence-based Protein Language Model (PLM) fine-tuned for the task of generating paired human variable heavy and light chain antibody sequences against targets of interest. We show that MAGE can generate novel and diverse antibody sequences with experimentally validated binding specificity against SARS-CoV-2, an emerging avian influenza H5N1, and respiratory syncytial virus A (RSV-A). MAGE represents a first-in-class model capable of designing human antibodies against multiple targets with no starting template.

Matching journals

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

1
Science
● 477 papers in training set
Top 0.4%
12.9%
Similar papers in this journal
2
mAbs
○ 32 papers in training set
Top 0.1%
12.5%
Similar papers in this journal
3
Science Advances
● 1243 papers in training set
Top 2%
8.9%
Similar papers in this journal
4
Nature Communications
○ 5641 papers in training set
Top 21%
7.9%
Similar papers in this journal
5
Proceedings of the National Academy of Sciences
● 2444 papers in training set
Top 8%
5.5%
Similar papers in this journal
6
Cell Reports
○ 1498 papers in training set
Top 8%
4.8%
50% of probability mass above
7
iScience
○ 1154 papers in training set
Top 4%
4.0%
8
Cell Systems
○ 201 papers in training set
Top 1%
3.4%
9
Nature Methods
○ 385 papers in training set
Top 3%
3.2%
10
Cell
○ 431 papers in training set
Top 4%
2.6%
11
Science Immunology
● 88 papers in training set
Top 1%
2.1%
12
Communications Biology
○ 993 papers in training set
Top 14%
1.7%
13
Antibody Therapeutics
◐ 16 papers in training set
Top 0.1%
1.7%
14
Cell Reports Methods
○ 165 papers in training set
Top 2%
1.5%
15
eLife
● 5828 papers in training set
Top 55%
1.4%
16
Frontiers in Immunology
○ 638 papers in training set
Top 7%
1.3%
17
Computational and Structural Biotechnology Journal
● 242 papers in training set
Top 5%
1.1%
18
PLOS Computational Biology
● 1863 papers in training set
Top 17%
1.1%
19
Cell Genomics
○ 172 papers in training set
Top 3%
1.1%
20
Cell Reports Medicine
○ 153 papers in training set
Top 3%
1.1%
21
Nucleic Acids Research
◐ 1281 papers in training set
Top 11%
1.1%
22
Advanced Science
○ 286 papers in training set
Top 7%
1.1%
23
Nature
○ 645 papers in training set
Top 9%
1.1%
24
Cell Host & Microbe
126 papers in training set
Top 2%
0.8%
25
Journal of Molecular Biology
○ 232 papers in training set
Top 4%
0.8%
26
Nature Machine Intelligence
○ 70 papers in training set
Top 2%
0.8%
27
ACS Central Science
● 71 papers in training set
Top 2%
0.6%
28
Scientific Reports
○ 3612 papers in training set
Top 76%
0.6%
29
Nature Biomedical Engineering
○ 47 papers in training set
Top 2%
0.6%
30
Nature Biotechnology
○ 172 papers in training set
Top 5%
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.