Using Supervised Machine Learning Algorithms to Screen Down Syndrome and Identify the Critical Protein Factors

Down syndrome (DS) is a genetic disorder caused by trisomy of all or part of the human chromosome 21. Since currently there is no cure for Down syndrome, the screening tests became the most efficient ways for DS prevention. Here, we used various supervised learning algorithms to build DS classification/screening models based on the protein/protein modification expression level of mice DS model Ts65Dn. Furthermore, we applied an adaptive boosted Decision Tree method to identify the most correlated and informative proteins factors that were associated with DS biological processes and pathways. Moreover, we improved the DS classification/screening models by using these selected DS related critical proteins. Finally, we used unsupervised learning algorithms to confirm the results we obtained above. These selected DS related proteins could be further used for protein-related and coding gene-related drugs developments.