Hui Tiffani, Secor Maxim, Ho Minh Ngoc, Bayaraa Nomindari, Lin Yu-Shan
Department of Chemistry, Tufts University, Medford, Massachusetts 02155, United States.
J Chem Inf Model. 2025 Feb 24;65(4):1837-1849. doi: 10.1021/acs.jcim.4c02102. Epub 2025 Feb 2.
Machine learning (ML) models have become increasingly popular for predicting and designing structures and properties of peptides and proteins. These ML models typically use peptides and proteins containing only canonical amino acids as the training data. Consequently, these models struggle to make accurate predictions for peptides and proteins containing new amino acids that are absent in the training data set (, noncanonical amino acids). One approach to improve the accuracy of the models is to collect more training data with the desired amino acids. However, this strategy is suboptimal as new data may not be easily attainable, and additional time is required to retrain the ML models. Alternatively, the extendibility of the ML models can be improved if the amino acid features used are representative and generalizable to the unseen amino acids. Herein, we develop amino acid features using molecular dynamics (MD) simulation results. Specifically, for a given amino acid, we perform MD simulation of its dipeptide to create features based on its backbone (ϕ, ψ) distributions and its electrostatic potentials. We demonstrate that these new features enable our ML models to more accurately predict the structural ensembles of cyclic peptides containing amino acids not present in the original training data set. For example, we build ML models to predict cyclic pentapeptide structures, with the training data set containing a library of 15 amino acids and the test data set containing the same 15-amino-acid library or an extended 50-amino-acid library. When using popular features such as Morgan fingerprints and MACCS keys to represent amino acids, the ML models achieve = 0.963 for structural predictions of test cyclic pentapeptides containing the same 15-amino-acid library. However, these models' performances decrease significantly to = 0.430 and = 0.508, respectively, when tasked to predict the structures of cyclic pentapeptides containing a library of 50 amino acids. On the other hand, the model using our backbone (ϕ, ψ) features outperforms those using Morgan fingerprints and MACCS keys, with = 0.700. Overall, instead of having to collect more training data, our new features enable predictions of peptide sequences containing amino acids not originally present in the training data set at the mere cost of performing new dipeptide simulations for the new amino acids.
机器学习（ML）模型在预测和设计肽及蛋白质的结构与性质方面越来越受欢迎。这些ML模型通常使用仅包含标准氨基酸的肽和蛋白质作为训练数据。因此，对于含有训练数据集中不存在的新氨基酸（即非标准氨基酸）的肽和蛋白质，这些模型难以做出准确预测。提高模型准确性的一种方法是收集更多含有目标氨基酸的训练数据。然而，这种策略并不理想，因为新数据可能不容易获得，且需要额外时间重新训练ML模型。或者，如果所使用的氨基酸特征具有代表性且能推广到未见过的氨基酸，则可以提高ML模型的可扩展性。在此，我们利用分子动力学（MD）模拟结果开发氨基酸特征。具体而言，对于给定的氨基酸，我们对其二肽进行MD模拟，以基于其主链（ϕ，ψ）分布及其静电势创建特征。我们证明，这些新特征使我们的ML模型能够更准确地预测含有原始训练数据集中不存在的氨基酸的环肽的结构集合。例如，我们构建ML模型来预测环五肽结构，训练数据集包含一个15种氨基酸的文库，测试数据集包含相同的15种氨基酸文库或扩展的50种氨基酸文库。当使用摩根指纹和MACCS键等常用特征来表示氨基酸时，ML模型对含有相同15种氨基酸文库的测试环五肽的结构预测得分为=0.963。然而，当任务是预测含有50种氨基酸文库的环五肽结构时，这些模型的性能分别显著下降至=0.430和=0.508。另一方面，使用我们的主链（ϕ，ψ）特征的模型优于使用摩根指纹和MACCS键的模型，得分为=0.700。总体而言，我们的新特征无需收集更多训练数据，只需为新氨基酸进行新的二肽模拟，就能预测含有训练数据集中原本不存在的氨基酸的肽序列。
