Conclusion

Knowledge of the tertiary structures of proteins is crucial to understanding protein function, but the traditional laboratory methods of determining these structures cannot keep up with demand. In response to this bottleneck, computer algorithms that can predict the structure of a protein based on its sequence have been developed and have become exponentially more accurate in recent years. The best algorithms today are almost as accurate as laboratory analyses like X-ray crystallography and NMR spectroscopy.

Since winning CASP14 in 2020, the AlphaFold 2 algorithm has captivated protein scientists with its unprecedented speed and its ability to accurately predict the structures of difficult molecules. Its multiple-chain specific extension, AlphaFold-Multimer, has received similar acclaim. But while these algorithms are free for public use, they require expensive, specialized computing resources to install and run, as well as an operator who is comfortable running complex command-line scripts. Furthermore, the open-source versions of AlphaFold 2 and AlphaFold-Multimer do not provide utilities for viewing the finished predictions, meaning that an additional application must be mastered for the analysis portion of the workflow.

DNASTAR has removed these obstacles by offering a way to use AlphaFold 2 and AlphaFold-Multimer from any Windows or Mac laptop or desktop computer. NovaFold AI and NovaFold AI-Multimer both use the intuitive Protean 3D interface for effortless setup and prediction status monitoring, as well as for viewing and analyzing the predicted models. The predictions themselves take place “behind the scenes” in the secure Amazon (AWS) cloud, freeing up computing resources for other tasks.