Machine Learning Enables Accurate Prediction of Asparagine Deamidation Probability and Rate

Abstract

The spontaneous conversion of asparagine residues to aspartic acid or iso-aspartic acid, via deamidation, is a major pathway of protein degradation and is often seriously disruptive to biological systems. Deamidation has been shown to negatively affect both in vitro stability and in vivo biological function of diverse classes of proteins. During protein therapeutics development, deamidation liabilities that are overlooked necessitate expensive and time-consuming remediation strategies, sometimes leading to termination of the project. In this paper, we apply machine learning to a large (n = 776) liquid chromatographytandem mass spectrometry (LC-MS/MS) dataset of monoclonal antibody peptides to create computational models for the posttranslational modification asparagine deamidation, using the random decision forest method. We show that our categorical model predicts antibody deamidation with nearly 5% increased accuracy and 0.2 MCC over the best currently available models. Surprisingly, our model also paces or outperforms advanced and conventional models on an independent non-antibody dataset. In addition to deamidation probability, we are able to accurately predict deamidation rate (R 2 = 0.963 and Q2 = 0.822), a capability with no peer in current models. This method should enable significant improvement in protein candidate selection, especially in biopharmaceutical development, and can be applied with similar accuracy to enzymes, monoclonal antibodies, next-generation formats, vaccine component antigens, and gene therapy vectors such as adenoassociated virus.

Engineering Stability, Viscosity, and Immunogenicity of Antibodies by Computational Design

“…25,[101][102][103][104][105][106][107][108][109][110][111] One of the most common degradation events is the chemical modification of Asn and Asp residues, which share a degradation pathway. 25 Many of the methods to predict such degradation are statisticalbased methods, and experimental data to derive such prediction models are either from in-house experiments 101,103,107,108,110 or from literature. 104,106,111 For example, to understand origins of Asn deamidation and Asp isomerization, Sydow et al 101 used mass spectrometry to experimentally characterize 37 antibodies that were subjected to forced degradation.…”

“…In another study, Yan et al 107 used 10 antibodies under both normal and stressed conditions, and experimentally characterized the Asn deamidation, leading to a decision tree model to predict the Asn deamidation probability from antibody structures. More recently, based on in-house LC-MS/MS experiments and literature information, Delmar et al 108 used machine learning to predict Asn deamidation probability and rate. The training set consisted of 776 Asn residues from 67 antibodies.…”

Antibody Disulfide Bond Reduction and Recovery during Biopharmaceutical Process Development - A Review

“…In their studies, mAb structure-based features, sequence-based features and dynamic-based features were extracted as input data, and mAb peptide mapping results on oxidation were extracted as output data for model training. Similar machine learning algorithms were also developed recently for mAb deamidation risks prediction (Delmar, Wang, Choi, Martins, & Mikhail, 2019;Jia & Sun, 2017). The success of machine learning in predicting the mAb biophysical and biochemical stabilities provides the possibilities of applying machine learning algorithm to predict mAb disulfide bond reduction risks.…”

In silico Techniques for Prospecting and Characterizing Monoclonal Antibodies

“…Therefore, a conformational study is essential to highlight the residues liable to the chemical change. Chemical stability is generally based on statistical analysis derived from experiments or databases available in the literature, although some computational methods are being used [96,[102][103][104][105][106][107][108]. Statistics-based methods depend on data from previous experiments and provide valuable information about the behavior of proteins, being excellent guides during the development of new antibodies.…”