Machine learning-augmented metaproteomics reveals metabolic adaptation of anammox bacteria under high nitrogen loading rate.

Anaerobic ammonium oxidation (anammox) is a biological process in which NH4+ serves as the electron donor and NO2- as the electron acceptor to produce N2. Anammox bacteria are the key agents responsible for this process, playing a crucial role in biological nitrogen removal from wastewater and the global nitrogen cycle. Investigating the functional characteristics of anammox bacteria will contribute to a better understanding of the anammox process and facilitate the development of anammox-based technologies. In this study, we focused on the anammox bacteria Candidatus Kuenenia, employing metaproteomics to identify 1193 encoded proteins, including a cohort of 866 functionally unannotated proteins. Here, functionally unannotated proteins specifically referred to proteins whose biological functions had not yet been clearly described or validated. To address this gap, we established a machine learning (ML) prediction algorithm based on protein sequences, employing seven algorithms including Support Vector Machine (SVM), Random Forest (RF), eXtreme Gradient Boosting (XGBoost), K-Nearest Neighbours (KNN), Logistic Regression (LR), Long Short-Term Memory (LSTM), and ProtBERT for functional classification. The RF model exhibited the best performance, achieving an accuracy of 82.6 %. Prediction results indicated that these uncharacterized proteins are primarily involved in metabolic pathways such as Amino acid metabolism, Metabolism of cofactors and vitamins, Carbohydrate metabolism, Energy metabolism, and Nucleotide metabolism. Further analysis revealed that under high nitrogen loading rate (NLR) conditions, the microbial community enhanced environmental adaptability by reprogramming metabolic resource allocation. Our integrated analysis suggests that Amino acid metabolism likely provides precursors for the synthesis of key functional enzymes such as hydroxylamine oxidase (HAO) and hydrazine synthase (HZS). Furthermore, the concurrent up-regulation of Energy metabolism and Metabolism of cofactors and vitamins is consistent with an increased demand for ATP and essential enzymes under high-load conditions. This study revealed the metabolic regulatory strategies of anammox bacteria under high NLR conditions at the protein level, providing a novel theoretical foundation for optimizing the operational efficiency and stress resistance regulation of the anammox process.