Analysis of Breast Cancer Subtypes Prediction Based on Alternative Splicing Disorders

Xu Peng,Wang Bing,Fang Gang,Shi Xiaolong,Liu Wenbin
DOI: https://doi.org/10.11999/JEIT190871
2020-01-01
Abstract:Alternative splicing is closely related to the occurrence and development of a variety of complex diseases, the emergence of various diseases including tumors is often accompanied by the occurrence of alternative splicing disorders. The existing analysis of breast cancer subtypes is mainly based on single splicing isoform, and the difference in the overall distribution of splicing isoforms caused by alternative splicing disorders among subtypes is not considered. Therefore, a prediction method of breast cancer subtypes based on alternative splicing disorders is proposed, which mainly uses Jensen-Shannon(JS) divergence to find genes with large differences in alternative splicing disorders between subtypes, then constructes Back Propagation(BP) neural network model to classify breast cancer subtypes. The results show that this method could not only effectively detect tumor heterogeneous molecules, but also had good identification results in the classification of breast cancer subtypes, with an average F1-score of 0.89, and could provide personalized drug recommendations for patients with breast cancer subtypes. This study will effectively promote the development of breast cancer subtypes based on alternative splicing disorders.
What problem does this paper attempt to address?