Polypeptide amino acid sequence generation model training and related products

• Main Authors: CAI LIFENG, ZHAO HUAJIAN, SONG GENGSHEN
• Format: Patent
• Language: Chinese; English
• Published: 15.03.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTEDFOR SPECIFIC APPLICATION FIELDS; PHYSICS

The invention provides polypeptide amino acid sequence generation model training and a related product. A specific embodiment of the method comprises the following steps: by utilizing an active polypeptide amino acid sequence which is combined with a target receptor and has activity, performing lemma segmentation on the active polypeptide amino acid sequence to obtain an active polypeptide lemma sequence; then respectively performing feature embedding representation on fragments of the active polypeptide lemma sequence and a corresponding shifted lemma sequence after shifting one lemma backwards, and respectively taking the fragments and the shifted lemma sequence as a sample polypeptide lemma feature sequence of sample feature data and a label lemma feature sequence expected to be output, and then taking the sample polypeptide lemma feature sequence as input data; and training a polypeptide amino acid sequence generation model by taking the tag lemma feature sequence as supervision data. Furthermore, the pol