The invention discloses a type 2
diabetes mellitus mouse model construction method combining
artificial intelligence prediction and multi-target knockout. Acquiring and standardizing mouse multi-
omics data and
phenotype baseline data to form a standardized modeling
data set; constructing an interaction map covering the functions of the
pancreatic beta cells on the basis of the data; screening to obtain a target combination and a target
phenotype prediction result, generating a construction instruction, and submitting the construction instruction to a controlled facility to execute multi-
gene knockout transformation, so as to obtain a transformed mouse and a transformation
receipt; acquiring actual measurement
phenotype data of the transformed mouse based on the transformation
receipt, and comparing the actual measurement phenotype data with a target phenotype prediction result to generate phenotype deviation measurement and a phenotype judgment result; and when the phenotype determination result is satisfied, registering the target combination, the prediction result, the actually measured phenotype data and the transformation
receipt to form the type-2
diabetes mellitus composite mouse model. According to the method, AI-driven multi-target genetic modification closed-
loop optimization is realized, and the efficiency and accuracy of complex
metabolic disease animal model construction are improved.