Discovery Of Anti-tumor Peptide Based On Artificial Intelligence Prediction Algorithm And Their Preliminary Activity Evaluation

Objective:To construct an artificial intelligence drug screening system for insect-derived peptides,and to screen and predict antitumor peptides.Methods:Paper based on a typical structure characteristics of antitumor activity of antimicrobial peptides,using known anti-cancer peptides training data set of artificial intelligence algorithm in test,and optimize the prediction effect is the best algorithm for virtual screening Uniprot database has the characteristics of the peptide,and found a number of lead compounds with anticancer potential.Subsequently,the 20 peptides with the highest score were selected and chemically synthesized using the solid-phase peptide synthesis technology based on Fmoc.The synthesized peptides were characterized by high performance liquid chromatography and mass spectrometry.Finally,the proliferation inhibitory activity of the related peptides on different tumor cells was evaluated by MTT assay.Results:1.After comparing the performance of the prediction algorithm,it is found that the optimized stochastic forest model has the best prediction effect.2.Virtual screening of article 20 of the highest rated polypeptide related physical and chemical properties and the spiral wheel figure analysis showed that most of these peptides is rich in lysine and arginine and other basic amino acid cation with amphiphilic structures,both have close fat sex and is hydrophilic,these all conform to have a lot of typical characteristics of antitumor activity of antimicrobial peptides.3.High performance liquid chromatography(HPLC)and mass spectrometry(MS)showed that the purity of 20 polypeptides was above 90%.4.The cell experiment results showed that polypeptide 14 had a significant inhibitory effect on the proliferation of B16,MDA-MB-231,Hep G2,cac O-2 and other 4 tumor cells,and polypeptide 7,12,and 13 had a certain inhibitory effect on the proliferation of B16.Conclusion:This study successfully identified a group of leading polypeptides with potential anticancer activity,providing new ideas and options for new drug development.However,the application scope and accuracy of these algorithms need to be improved.