31 May 2022 | Tuesday | News
Figure 1. Predicted structure of Monkeypox Virus Congo Strain: Virus termination factor large subunit (Q3I8R6); Figure 2. PointSite Prediction; Figure 3. Structure-blast Prediction
Zelixir Biotech announces it has predicted more than 600 structures and added functional annotations of proteins from monkeypox virus proteomes amid the recent emergence of monkeypox cases in Europe and the United States. The predicted structures are expected to help experts worldwide to develop structure-based drug design and monkeypox vaccines.
The predicted protein structures are AI-generated by AlphaFold2-Batch (originated from the DeepMind AlphaFold2, ), the company's self-developed MPI-based structure prediction pipeline with heterogeneous parallelism powered by supercomputing infrastructures. The predictions are based on the complete sequence of the genome of the epidemic strain in May 2022 (ID: ON563414.2) (Ref. 1) , the whole proteome of the 2018 West African (WA) virus strain (Monkeypox virus isolate MPXV-UK_P3, ID: MT903345.1) (Ref.2), and the whole proteome of the 1996 Congo basin (CB) virus strain (Monkeypox virus strain Zaire-96-I-16, ID: NC_003310) (Ref. 3).
AlphaFold2-Batch, which is available on the company's platform "ZCloud", has largely boosted both the efficiency and the precision of the predictions by nearly two orders of magnitude acceleration compared to the original AlphaFold2 pipeline, generating results without any compromise of accuracies.
The precision of the AI-generated result is also supported by profound and up-to-date research by Zelixir. The published results provide functional annotations and analyses of the proteome from the latest west African strain compared with its counterpart, the 1996 Congo Basin strain.
Using the PointSite Algorithm developed by the Zelixir, the team provides profound annotations for all the 600+ predicted structures and managed to label the small molecules binding regions with high confidence for the three-dimensional structure of each protein. Meanwhile, experimentally solved structures with high similarity to monkeypox proteins are vetted through the company's structure-alignment algorithm against a subset of Protein Data Bank database (PDB70).
All the detailed forecasts and the corresponding functional predictions are available on its website (https://www.zelixir.com/Monkeypox/index.html).
"We hope our work can accelerate the development of monkeypox vaccines and drug discovery. We will keep updating and publishing our new research findings as our research progresses." said Dr. Sheng Wang, the CEO of Zelixir Biotech.