Designing Feature-Controlled Humanoid Antibody Discovery Libraries Using Generative Adversarial Networks

Abstract

We demonstrate the use of a Generative Adversarial Network (GAN), trained from a set of over 400,000 light and heavy chain human antibody sequences, to learn the rules of human antibody formation. The resulting model surpasses common in silico techniques by capturing residue diversity throughout the variable region, and is capable of generating extremely large, diverse libraries of novel antibodies that mimic somatically hypermutated human repertoire response. This method permits us to rationally design de novo humanoid antibody libraries with explicit control over various properties of our discovery library. Through transfer learning, we are able to bias the GAN to generate molecules with key properties of interest such as improved stability and developability, lower predicted MHC Class II binding, and specific complementarity-determining region (CDR) characteristics. These approaches also provide a mechanism to better study the complex relationships between antibody sequence and molecular behavior, both in vitro and in vivo . We validate our method by successfully expressing a proof-of-concept library of nearly 100,000 GAN-generated antibodies via phage display. We present the sequences and homology-model structures of example generated antibodies expressed in stable CHO pools and evaluated across multiple biophysical properties. The creation of discovery libraries using our in silico approach allows for the control of pharmaceutical properties such that these therapeutic antibodies can provide a more rapid and cost-effective response to biological threats.

Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences

“…Bepler and Berger ( 70 ) pretrained LSTMs on protein sequences, adding supervision from contacts to produce embeddings. Subsequent to our preprint, related works have built on its exploration of protein sequence modeling, exploring generative models ( 71 , 72 ), internal representations of Transformers ( 73 ), and applications of representation learning and generative modeling such as classification ( 74 , 75 ), mutational effect prediction ( 80 ), and design of sequences ( 76 – 78 ).…”

Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences

“…Bepler & Berger (2019) pre-trained LSTMs on protein sequences, adding supervision from contacts to produce embeddings. Subsequent to our preprint, related works have built on its exploration of protein sequence modeling, exploring generative models (Riesselman et al, 2019;Madani et al, 2020), internal representations of Transformers (Vig et al, 2020), and applications of representation learning and generative modeling such as classification (Elnaggar et al, 2019;Strodthoff et al, 2020), mutational effect prediction (Luo et al, 2020), and design of sequences (Repecka et al, 2019;Hawkins-Hooker et al, 2020;Amimeur et al, 2020).…”

A compact vocabulary of paratope-epitope interactions enables predictability of antibody-antigen binding

“…Monoclonal antibody discovery is predominantly performed using synthetic antibody libraries. The number of developable hits of such libraries may be increased by tuning sequence diversity toward the interaction motifs (and their corresponding sequential bias) discovered here (Amimeur et al, 2020;Chen et al, 2020). Relatedly, engineering-driven computational optimization of antibody-antigen binding, as well as docking algorithms, might benefit from incorporating interaction-motif-based heuristics (Baran et al, 2017;Krawczyk et al, 2013;Kuroda and Gray, 2016;Mason et al, 2019;Sivasubramanian et al, 2009;Weitzner and Gray, 2017).…”