Integrative Similarity analysis, Docking and Machine Learning models for identifying new Zika NS5 hits guided by Dengue NS5 inhibitors

The Zika virus (ZIKV) causes epidemics around the world1, but despite its severe consequences, there are no antivirals to combat infection. The NS5 protein, methyltransferase (Mtase) and RNA polymerase (RdRP) domains, has an essential role in viral RNA synthesis2. In this work, we searched for Dengue virus (DENV)3 NS5 inhibitors in PubChem and ChEMBL databases. We performed an integrative similarity analysis of DENV and ZIKV NS5 proteins, docking of DENV known inhibitors on ZIKV NS5 sites and machine learning (ML) models, to prioritize the best hits. We found 156 compounds reported as DENV NS5 inhibitors, that were docked in the ZIKV NS5 sites and scored by ML models. Twenty-two compounds were selected. Then, we performed a similarity-based search with the hits in a commercial database and screened the similar compounds using the docking and ML filters. The 67 virtual hits for ZIKV NS5 will be validated by cell-based ZIKV assays.