An extension of Wang protein design model using Blosum62 substitution matrix
Rahmani, A.; Zare-Mirakabad, F.
Show abstract
Humans life depends on the functionality of molecules in the body. One of these essential molecules is the protein that plays a vital role in our life, such that its malfunction can cause severe damages. Such roles make protein structure and its functionality necessary to understand. One of the problems that help us understand the relation between protein structure is the well-known protein design problem which attempts to find an amino acid sequence that can fold into a desired tertiary structure. However, despite having an acceptable accuracy in protein design, this accuracy is an identical percentage of amino acid retrieving. At the same time, it is well-known that amino acids can replace each other in evolution while the function and structure of protein stay the same. Thus the designed sequence does not have the opportunity to be close to the target in the evolutionary aspect. This paper presents an extension to Wangs deep learning model, which uses evolutionary information in the Blosum62 substitution matrix to take amino acid replacement probability into account while designing a sequence.
Matching journals
The top 6 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Rprot-Vec: A deep learning approach for fast protein structure similarity calculation 98%
- Multi-Head Attention-based U-Nets for Predicting Protein Domain Boundaries Using 1D Sequence Features and 2D Distance Maps 98%
- Binding affinity prediction for protein-ligand complex using deep attention mechanism based on intermolecular interactions 96%
Similar papers in this journal
- Improving prediction of drug-target interactions based on fusing multiple features with data balancing and feature selection techniques 96%
- Antivirals for Monkeypox Virus: Proposing an Effective Machine/Deep Learning Framework 96%
- Predicting Adverse Drug Effects: A Heterogeneous Graph Convolution Network with a Multi-layer Perceptron Approach 96%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.