Abstract
Molecular evolution is an important step in the development of therapeutic antibodies. However, the current method of affinity maturation is overly costly and labor-intensive because of the repetitive mutation experiments needed to adequately explore sequence space. Here, we employed a long short term memory network (LSTM)—a widely used deep generative model—based sequence generation and prioritization procedure to efficiently discover antibody sequences with higher affinity. We applied our method to the affinity maturation of antibodies against kynurenine, which is a metabolite related to the niacin synthesis pathway. Kynurenine binding sequences were enriched through phage display panning using a kynurenine-binding oriented human synthetic Fab library. We defined binding antibodies using a sequence repertoire from the NGS data to train the LSTM model. We confirmed that likelihood of generated sequences from a trained LSTM correlated well with binding affinity. The affinity of generated sequences are over 1800-fold higher than that of the parental clone. Moreover, compared to frequency based screening using the same dataset, our machine learning approach generated sequences with greater affinity.
Highlights
Molecular evolution is an important step in the development of therapeutic antibodies
We perform panning against an antigen, e.g., kynurenine (Phage display panning), and we get a large-scale enriched antibody sequence by next-generation sequencing (NGS) (NGS)
We begin generating virtual sequences that mimic enriched sequences from B based on the long short term memory network (LSTM) model (Sequence generation)
Summary
Molecular evolution is an important step in the development of therapeutic antibodies. We employed a long short term memory network (LSTM)—a widely used deep generative model—based sequence generation and prioritization procedure to efficiently discover antibody sequences with higher affinity. We defined binding antibodies using a sequence repertoire from the NGS data to train the LSTM model. We confirmed that likelihood of generated sequences from a trained LSTM correlated well with binding affinity. There is a clear demand to find measures other than frequency when utilizing NGS derived sequences to discover promising candidates. To address these issues, we employed a long short term memory network (LSTM)-based sequence generation and prioritization procedure to efficiently discover sequences with higher affinity. The development of more powerful screening methods for obtaining high-affinity antibodies is desired
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have