is structural information a must in intermolecular binding affinity calculation Is Protein–protein binding affinity prediction from amino acid sequence adequately accurate

Structural information is not always a must in intermolecular binding affinity calculation, as there are computational methods available that can predict binding affinities solely from sequence information. These methods, known as "sequence-based" or "sequence-only" methods, utilize various algorithms and machine learning techniques to predict binding affinities based on amino acid sequence alone.

However, it is important to note that the accuracy of these sequence-based methods can vary depending on the specific protein-protein interaction and the availability of high-quality training data. In some cases, incorporating structural information such as protein structures or binding site information can improve the accuracy of the predictions.

Overall, while sequence-based methods can provide useful predictions, they may not always be adequately accurate on their own, especially for complex protein-protein interactions. Therefore, combining sequence-based predictions with structural information can enhance the accuracy of binding affinity predictions.