ILMCNet: A Deep Neural Network Model That Uses PLM to Process Features and Employs CRF to Predict Protein Secondary...

Area of Science:

• Computational biology
• Bioinformatics
• Structural bioinformatics

Background:

• Protein secondary structure prediction (PSSP) is crucial for understanding protein function and disease.
• Integrating diverse amino acid sequence features enhances PSSP accuracy.

Purpose of the Study:

• To develop a deep neural network model for PSSP that incorporates evolutionary and positional information.
• To account for interdependencies between secondary structures during prediction.

Main Methods:

• ILMCNet model using a protein language model (PLM) and Conditional Random Field (CRF).
• Incorporation of positional encoding and a hybrid network (Transformer Encoder, CNN, BiLSTM) for feature extraction.
• CRF layer to model secondary structure interdependencies.

Main Results:

• ILMCNet achieved superior prediction performance on benchmark datasets (CB513, TS115, NEW364, CASP11, CASP12).
• The model's performance surpassed that of comparable existing approaches.

Conclusions:

• The proposed ILMCNet offers a novel and effective approach to PSSP.
• This method is expected to advance protein tertiary structure prediction and other related fields.