SDAB_GUI: Graphical User Interface for Novel Single-Domain Antibody Generation Using Generative Pre-Trained Transformers and Recurrent Neural Networks

This report is an update to the first version of the submitted work, Graphical User Interface for Novel Protein Generation using ProtGPT: Version 1, which detailed the Year 4 efforts of the Naval Research Laboratory (NRL) toward the algorithm development for optimization of biologic medical countermeasures. The directed modification of antibody proteins and the foundational implications of reengineering and synthesizing of novel, small molecules with desired properties are further developed in modern biotechnology to combat antibiotic resistance. Specifically, the expensive and production delays of novel single-domain antibodies (sdAbs) pose the challenge in antibiotic discovery. This, however, can be mitigated with fast, efficient, and automated sdAb sequence generation ready to be synthesized for potential therapeutic use. Thus, we developed SDAB_GUI a ready-to-use deployable software with a graphical user interface (GUI) using an open-source, interactive application which stemmed from a previously reported prototype. SDAB_GUI incorporates generative pre-trained transformer (GPT) and recurrent neural network (RNN) models for sdAb generation and enables users to follow a set of cascaded steps for the input, exploration, and generation of new sequences based on an existing sdAb dataset. Summary statistics of the input and output sdAb sequence dataset, exploratory analyses, and the proteins physicochemical characteristics can also be interactively employed. Validation measures are included along the sequence generation such as redundancies or antibody numbering annotations to produce proper sdAb characteristics. This automated process allows the generation of novel sdAb sequences that can subsequently be evaluated for improvements to a specifically desired biophysical property.This document also serves as the reference material for the SDAB GUI software.