Drugs and nutraceuticals for SARS-CoV infection. An exercise of repurposing based on AI.

ChEMBL DB contains >15M experimental records with more than 1.9M molecules.
Machine learning models built on this basis have identified  66k potentially active compounds  against SARS-CoV. 
Among these 1.9M compounds there are 37k known substances with identified names and pharmaceutical, agricultural or nutritional use, 4k out of them identified as potentially actives against coronaviruses applying a sarsScore cutoff = 4.

Activity distribution histogram of known ChEMBL substances.

drTarget team decided to study the value of these substances by a combined approach.
  • Drug-repurposing by studying all interactions of selected active drugs with protein targets recorded in ChEMBL DB. 
  • Nutraceuticals analysis by examining the contents in food of substances identified as active. Nutritional contents of foods taken from FOODB


Eventually, nutraceuticals appear in ChEMBL interactions analysis. 

Drug-repusposing schema. 

Some nutraceuticals for SARS-CoV infection can be found in a broad spectrum of foods. 

We can depict a drugs-proteins interactions graph from All ChEMBL records for predicted active drugs and thus associate antiviral activity to different mechanisms of action. 

And, similarly, we can identify the foods where potentially therapeutic nutrients are more abundant by using the downloadable info provided by FOODB