A de novo sequencing method for selenopeptides based on a fully convolutional attention mechanism neural network
CN119479810A · Shaanxi University of Science and Technology · priority 2024
| Patent | CN119479810A ↗ |
| Status | Pending |
| Priority date | 2024-10-30 |
| Filed | 2024-10-30 |
| Published | 2025-02-18 |
| Assignee | Shaanxi University of Science and Technology |
Abstract
The invention provides a selenium peptide de-header sequencing method based on a full convolution attention mechanism neural network, which belongs to the selenium-containing natural polypeptide identification detection field, and comprises the steps of taking specific seleno-amino acid corresponding to each characteristic fragment in all selenium peptides as initial amino acid, then predicting the next symbol in the N/C end direction in sequence, taking the highest probability symbol as the symbol of the corresponding position, learning the amino acid characteristics of the selenium peptide N, C end in a known selenium peptide sequence by using the full convolution attention mechanism neural network, then carrying out characteristic fusion, standardizing to 0 to 1, taking all identified selenium peptide sequences as input layers, carrying out data characteristic separation by using an activation function of a neural network hidden layer, calculating each separation characteristic, obtaining the probability of the corresponding selenium peptide sequence as the natural selenium peptide, removing the unmatched selenium peptide sequence, outputting the identified natural selenium peptide sequence by an output layer, optimizing the weight and bias of a model, inputting all identified selenium peptide sequences into the optimized model, and inputting the selenium peptide sequences which are correspondingly identified by the tested sample into the sequencing model when the error between the selenium peptide sequences is unchanged, so as to obtain the natural selenium peptide sequence to be tested.
Why it is here
Shaanxi University of Science and Technology: de novo sequencing of selenopeptides with a fully convolutional attention network.
Inventors (3)
贾玮, 张�荣, 武希璇
An inventor links to an author page only where the same person is evident, never on a name alone.