A Deep Retrieval-Enhanced Meta-Learning Framework for Enzyme Optimum pH Prediction

The potential of hydrogen (pH) influences the function of the enzyme. Measuring or predicting the optimal pH (pH_opt) at which enzymes exhibit maximal catalytic activity is crucial for enzyme design and application. The rapid development of enzyme mining and de novo design has produced a large number of new enzymes, making it impractical to measure their pHopt in the wet laboratory. Consequently, in-silico computational approaches such as machine learning and deep learning models, which offer pH prediction at minimal cost, have attracted considerable interest. This work presents Venus-DREAM, an enzyme pH_opt prediction model based on the <i>k</i>NN algorithm and few-shot learning, which achieves state-of-the-art accuracy in pHopt prediction. Venus-DREAM regards the pH_opt prediction of an enzyme as a few-shot learning task: learning from the <i>k</i>-closest labeled enzymes to predict the pH_opt of the target enzyme. The value of <i>k</i> is determined by the optimal <i>k</i>-value of the <i>k</i>NN regression algorithm. And the distance between two enzymes is defined as the cosine similarity of their mean-pooled embeddings obtained from protein language models (PLMs). The few-shot learner is based on the Reptile algorithm, which first adapts to the <i>k</i>-nearest labeled enzymes to create a specialized model for the target enzyme and then predicts its pH_opt. This efficient method enables high-throughput virtual exploration of protein space, facilitating the identification of sequences with the desired pH_opt ranges in a high-throughput manner. Moreover, our method can be easily adapted in other protein function prediction tasks.