This website hosts a dual-functional anticancer peptide (ACP) predictor. The first function predicts whether an input peptide is an ACP, while the second classifies the anticancer activity type of confirmed ACPs. As emerging candidate molecules for cancer therapy, ACPs offer new hope due to their high efficacy and low propensity for inducing drug resistance. However, existing ACP identification methods primarily rely on peptide sequence features while neglecting spatial structural characteristics, and few approaches can simultaneously predict ACP functionality. To address these limitations, we propose Multi-ACPNet, a dual-functional predictor capable of both ACP identification and anticancer activity classification. The framework begins by employing an LSTM-Causal CNN hybrid module to learn ESM C features, capturing both global and local sequence correlations. Subsequently, the multi-scale sequence features are combined with handcrafted features and fed into a graph convolutional network (GCN) with residual connections. By constructing multi-scale neighborhoods and utilizing a weighted fusion mechanism, the model dynamically captures both local and long-range structural dependencies among amino acids.
For researchers, Multi-ACPNet accelerates the discovery of novel ACPs and their therapeutic potential. Try it now by submitting your peptide sequence!