WLLP: A weighted reconstruction-based linear label propagation algorithm for predicting potential therapeutic agents...

Area of Science:

• Computational biology and bioinformatics
• Drug discovery and development
• Infectious disease modeling

Background:

• The COVID-19 pandemic presents significant global health and economic challenges.
• Traditional drug and vaccine development is resource-intensive (labor, cost, time).
• Drug repositioning, aided by data science, offers a promising avenue for rapid therapeutic development.

Purpose of the Study:

• To propose a novel computational model for predicting effective drug regimens against SARS-CoV-2.
• To leverage drug repositioning strategies to accelerate the identification of COVID-19 treatments.
• To evaluate the performance of the proposed model against existing methods.

Main Methods:

• Development of a weighted reconstruction-based linear label propagation algorithm (WLLP).
• Utilized the human drug virus database (HDVD) for model training and prediction.
• Employed a weighted K-nearest neighbors method for network preprocessing and label propagation.

Main Results:

• WLLP demonstrated excellent predictive performance, achieving an AUC of 0.8828 ± 0.0037.
• The model outperformed four other comparative algorithms in predicting drug-virus associations.
• Case studies successfully identified potential drug regimens for COVID-19 treatment.

Conclusions:

• The WLLP algorithm is a highly effective tool for drug repositioning in the context of viral pandemics.
• This computational approach can significantly expedite the discovery of potential treatments for diseases like COVID-19.
• WLLP offers a valuable strategy to reduce the time and cost associated with developing new therapeutic options.