Welcome to FSL-Kla

As a novel lactate-derived post-translational modification ( PTM), lysine lactylation ( Kla) was discovered, which belongs to metabolite-derived PTMs similar with Lys acetylation ( Kac) ( Zhang et al., 2019). Biochemically, Kac introduce a small acetyl group on the ε amine group of the Lys residue, with a mass of 42.0106 Daltons (Da) ( Sabari et al., 2016). Kla attaches a lactyl group to the ε amino group of a lysine residue, with a much larger mass of 72.021 Da (Zhang et al., 2019). Similar to Kac, Kla occurs in both histone and non-histone proteins, and faithfully orchestrates numerous biological processes, such as signal transduction, metabolism and inflammatory responses (Zhang et al., 2019; Irizarry-Caro et al., 2019). In addition, dysregulation of lactylation contributes to tumorigenesis ( Yu et al., 2021). Kla represents a typical non-metabolic role for lactate and illuminate a new avenue to study the diverse physiological functions of lactate. Although the biological importance of protein lactylation has been gradually recognized in recent years, its underlying mechanisms are still largely unclear.

Identification of Kla substrates with their exact sites is crucial for revealing the molecular mechanisms of lactylation. In contrast with labor-intensive and time-consuming experimental approaches, computational prediction of Kla could provide convenience and increased speed, but is still lacking. In this work, although current identified Kla sites are limited, we constructed the first Kla benchmark dataset and utilize an ensemble deep learning-based few-shot learning ( FSL) approach to leverage the power of small datasets to avoid the impact of imbalance and overfitting. Furthermore, we conducted a comprehensive survey of performances by combining 8 sequence-based features, physicochemical properties and 3 structure-based features. Then, we present our newly designed predictor, FSL-Kla, which achieved a at least 16.2% improvement of the area under the curve ( AUC) value ( 0.889 versus 0.765) for the perdition of Kla sites in silico. The webserver for FSL-Kla freely accessible for academic research at http://kla.zbiolab.cn/.