Protein language models enable prediction of polyreactivity of monospecific, bispecific, and heavy-chain-only antibodies.

Q2 Medicine Antibody Therapeutics Pub Date : 2024-05-30 eCollection Date: 2024-07-01 DOI:10.1093/abt/tbae012
Xin Yu, Kostika Vangjeli, Anusha Prakash, Meha Chhaya, Samantha J Stanley, Noah Cohen, Lili Huang
{"title":"Protein language models enable prediction of polyreactivity of monospecific, bispecific, and heavy-chain-only antibodies.","authors":"Xin Yu, Kostika Vangjeli, Anusha Prakash, Meha Chhaya, Samantha J Stanley, Noah Cohen, Lili Huang","doi":"10.1093/abt/tbae012","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>Early assessment of antibody off-target binding is essential for mitigating developability risks such as fast clearance, reduced efficacy, toxicity, and immunogenicity. The baculovirus particle (BVP) binding assay has been widely utilized to evaluate polyreactivity of antibodies. As a complementary approach, computational prediction of polyreactivity is desirable for counter-screening antibodies from <i>in silico</i> discovery campaigns. However, there is a lack of such models.</p><p><strong>Methods: </strong>Herein, we present the development of an ensemble of three deep learning models based on two pan-protein foundational protein language models (ESM2 and ProtT5) and an antibody-specific protein language model (PLM) (Antiberty). These models were trained in a transfer learning network to predict the outcomes in the BVP assay and the bovine serum albumin binding assay, which was developed as a complement to the BVP assay. The training was conducted on a large dataset of antibody sequences augmented with experimental conditions, which were collected through a highly efficient application system.</p><p><strong>Results: </strong>The resulting models demonstrated robust performance on canonical mAbs (monospecific with heavy and light chain), bispecific Abs, and single-domain Fc (VHH-Fc). PLMs outperformed a model built using molecular descriptors calculated from AlphaFold 2 predicted structures. Embeddings from the antibody-specific and foundational PLMs resulted in similar performance.</p><p><strong>Conclusion: </strong>To our knowledge, this represents the first application of PLMs to predict assay data on bispecifics and VHH-Fcs.</p>","PeriodicalId":36655,"journal":{"name":"Antibody Therapeutics","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11259759/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Antibody Therapeutics","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1093/abt/tbae012","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2024/7/1 0:00:00","PubModel":"eCollection","JCR":"Q2","JCRName":"Medicine","Score":null,"Total":0}
引用次数: 0

Abstract

Background: Early assessment of antibody off-target binding is essential for mitigating developability risks such as fast clearance, reduced efficacy, toxicity, and immunogenicity. The baculovirus particle (BVP) binding assay has been widely utilized to evaluate polyreactivity of antibodies. As a complementary approach, computational prediction of polyreactivity is desirable for counter-screening antibodies from in silico discovery campaigns. However, there is a lack of such models.

Methods: Herein, we present the development of an ensemble of three deep learning models based on two pan-protein foundational protein language models (ESM2 and ProtT5) and an antibody-specific protein language model (PLM) (Antiberty). These models were trained in a transfer learning network to predict the outcomes in the BVP assay and the bovine serum albumin binding assay, which was developed as a complement to the BVP assay. The training was conducted on a large dataset of antibody sequences augmented with experimental conditions, which were collected through a highly efficient application system.

Results: The resulting models demonstrated robust performance on canonical mAbs (monospecific with heavy and light chain), bispecific Abs, and single-domain Fc (VHH-Fc). PLMs outperformed a model built using molecular descriptors calculated from AlphaFold 2 predicted structures. Embeddings from the antibody-specific and foundational PLMs resulted in similar performance.

Conclusion: To our knowledge, this represents the first application of PLMs to predict assay data on bispecifics and VHH-Fcs.

查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
蛋白质语言模型可预测单特异性、双特异性和纯重链抗体的多反应性。
背景:抗体脱靶结合的早期评估对于降低开发风险(如快速清除、疗效降低、毒性和免疫原性)至关重要。杆状病毒颗粒(BVP)结合试验已被广泛用于评估抗体的多反应性。作为一种补充方法,多反应性的计算预测是反筛选硅学发现活动中抗体的理想方法。方法:在此,我们基于两个泛蛋白质基础蛋白质语言模型(ESM2 和 ProtT5)和一个抗体特异性蛋白质语言模型(PLM)(Antiberty),开发了三个深度学习模型的集合。在迁移学习网络中对这些模型进行了训练,以预测 BVP 检测和牛血清白蛋白结合检测的结果,牛血清白蛋白结合检测是作为 BVP 检测的补充而开发的。训练是在一个大型的抗体序列数据集上进行的,该数据集通过一个高效的应用系统收集了大量的实验条件:结果:由此产生的模型在典型 mAbs(重链和轻链的单特异性抗体)、双特异性抗体和单域 Fc(VHH-Fc)上表现出强大的性能。PLM 的表现优于使用 AlphaFold 2 预测结构计算的分子描述符建立的模型。来自抗体特异性和基础 PLM 的嵌入结果具有相似的性能:据我们所知,这是首次应用 PLM 预测双特异性抗体和 VHH-Fcs 的检测数据。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
Antibody Therapeutics
Antibody Therapeutics Medicine-Immunology and Allergy
CiteScore
8.70
自引率
0.00%
发文量
30
审稿时长
8 weeks
期刊最新文献
AI-based antibody discovery platform identifies novel, diverse, and pharmacologically active therapeutic antibodies against multiple SARS-CoV-2 strains. FcRider: a recombinant Fc nanoparticle with endogenous adjuvant activities for hybrid immunization. A pan-allelic human SIRPα-blocking antibody, ES004-B5, promotes tumor killing by enhancing macrophage phagocytosis and subsequently inducing an effective T-cell response. Correction to: A case study of a bispecific antibody manufacturability assessment and optimization during discovery stage and its implications. The process using a synthetic library that generates multiple diverse human single domain antibodies.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1