Thursday, December 2 2021
2:00am - 4:00am
Masters Presentation
Improving Omics-Phenotype Network Detection with Modified PageRank Method

Chronic obstructive pulmonary disease (COPD) is the fourth leading cause of death in the United States. Although COPD occurs predominantly in smokers, it is unknown why only a minority of smokers develops chronic airflow limitation. Recently, a canonical correlation analysis-based method (i.e., SmCCNet) was developed to identify complex relationships associated with lung function and the forced expiratory volume (FEV1). However, this method is not efficient as it uses a simple hierarchical graph clustering approach for detection of significant subnetworks, which removes important subnetworks by biasing toward finding compact clusters/subnets. This study aims to improve the SmCCNet subnetwork detection using a modified PageRank method. To evaluate the subnets, correlation and conductance measures were used. We applied these metrics in sequential and simultaneous ways. The highest obtained correlation was 0.35(p-value = 1.22e-29) which is about two times more than the former hierarchical results. The proposed subnetwork detection method assists in the discovery of novel biomarkers for COPD treatments.
Speaker:Mesbah Najafi
Affiliation:
Location:zoom


Download as iCalendar

Done