A parallel model of DenseCNN and ordered‐neuron LSTM for generic and species‐specific succinylation site prediction

Abstract

Lysine succinylation (Ksucc) regulates various metabolic processes, participates in vital life processes, and is involved in the occurrence and development of numerous diseases. Accurate recognition of succinylation sites can reveal underlying functional mechanisms and pathogenesis. However, most remain undetected. Moreover, a deep learning architecture focusing on generic and species‐specific predictions is still lacking. Thus, we proposed a deep learning‐based framework named Deep‐Ksucc, combining a dense convolutional network and ordered‐neuron long short‐term memory in parallel, which took the cascading characteristics of sequence information and physicochemical properties as the input. The results of the generic and species‐specific predictions indicated that Deep‐Ksucc can identify sequence patterns of different organisms and recognize plenty of succinylation sites. The case study showed that Deep‐Ksucc can serve as a reliable tool for biology verification and computer‐aided recognition of succinylation sites.

PD-BertEDL: An Ensemble Deep Learning Method Using BERT and Multivariate Representation to Predict Peptide Detectability

“…For Hybrid and PD-BertEDL, the difference is only in the final classification method. The MCC value of PD-BertEDL increases by 2.64% compared with Hybrid, and the MCC value can be used to measure the classification quality of a binary classifier [ 41 ]. This indicates that in the prediction of peptide detectability, after using different models to automatically obtain corresponding high-level features from specific information, ensemble classification method is more helpful to improve the prediction performance of peptide detectability than mixed model classification method.…”

“…From Table 4 , we know that compared with other predictors, the Sn, ACC, and MCC values of the proposed PD-BertEDL method are 92.38%, 82.58%, and 66.52%, respectively, but the Sp value is relatively low. Sn and Sp are antagonistic, and Sn value represents the percentage of data predicted as positive cases in all positive cases [ 41 ]. In protein identification analysis and peptide detection, it is important to know which theory digested peptides can be identified.…”

“…One-of-21 encoding was used to encode sequence information of the peptide, which is a discrete representation of value 1 at the index corresponding to the amino acid in the peptide and 0 at all other positions [ 41 ]. For example, the one-of-21 encoding of a sequence fragment ‘KFVICHLKGK’ is [[001000000000][000001000000]……[000000010000]].…”