AIkido Pharma provided and update on the use of Machine Learning in support of its antiviral platform with the University of Maryland Baltimore School of Medicine. 

The goal for the project is the identification and optimization of anti-viral compounds that inhibit viral replication by targeting a protein complex that degrades RNA at the cellular level.  To support this goal, physics-based machine learning (ML) is being applied by SilcsBio LLC to accelerate the discovery of broad-spectrum antivirals.

These efforts are targeting the human SKI complex that is involved in the replication of RNA viruses from which multiple drug candidates have been identified and shown to be effective against SARS-CoV-2 as well as other coronaviruses, including MERS-CoV, Influenza viruses and the filoviruses Ebola and Marburg. Ongoing efforts involve further development of those drug candidates using the SILCS data-driven ML ligand optimization approach in conjunction with medicinal chemistry, biophysical characterization, and cell- and animal-based antiviral experimental evaluation. These collaborative efforts will yield novel chemical entities to be considered for investigational new drug (IND) status and clinical trials leading to therapeutic agents poised to take on the next global pandemic.

Alex MacKerell, Grollman-Glick Professor of Pharmaceutical Sciences and CSO, SilcsBio, stated, "SilcsBio's machine learning algorithm combines physics-based simulations with a data-driven machine learning approach to iteratively improve the predictability of the ligand optimization process. In addition, our deep graph-based deep neural network allows screening of more than a billion compounds in a matter of days, allowing discovery of novel compounds much quicker than traditional approaches.  Towards addressing the global Covid-19 pandemic SilcsBio, LLC has partnered with AIkido Pharmaceuticals by applying its proprietary technology towards the development of novel broad-spectrum antivirals that will also be effective towards other viruses, including Influenza, Ebola and Marburg."

Prof. MacKerell went on to say, "The success of these drug development efforts are based on the power of the SilcsBio technology to unlock hidden binding hotspots on novel antiviral drug targets, the ability to identify sites on those targets that alter the interactions between novel interacting antiviral proteins and the rapid development of optimal and effective drug leads against those targets that will define the next generation of novel antiviral agents."