Metabolite-disease association prediction algorithm combining DeepWalk and random forest

Identifying the association between metabolites and diseases will help us understand pathogenesis of diseases, which has great significance in diagnosing treating diseases. However, traditional biometric methods are time consuming expensive. Accordingly, we propose a new metabolite-disease prediction algorithm based on DeepWalk random forest (DWRF), consists following key steps: First, semantic similarity information entropy integrated as final disease similarity. Similarly, molecular fingerprint metabolite Then, is used to extract features network metabolite-gene associations. Finally, employed infer The experimental results show that DWRF good performances terms area under curve value, leave-one-out cross-validation, five-fold cross-validation. Case studies also indicate reliable performance prediction.