Four Steps in Protein Structure Prediction
Whether you’re studying a single protein, an antibody-antigen complex, or a multi-component biological assembly, AI structure prediction generally follows the same four-step workflow.
Step 1 – Obtain an amino acid sequence
Every prediction begins with one or more protein sequences.
Sequences may be downloaded from public databases such as UniProt, the Protein Data Bank (PDB), and NCBI Protein. They can also be translated from newly sequenced DNA, or designed synthetically for protein engineering applications. Depending on the prediction method, a project may contain a single sequence or multiple interacting protein chains.
Step 2 – Choose a suitable prediction algorithm
The next step is selecting the prediction method that best matches your scientific question. An ever-increasing number of algorithms are available, and some of these are discussed in detail in Chapter 2.
Choosing the right algorithm can improve both prediction quality and turnaround time. Many researchers also compare results generated by multiple algorithms to gain additional confidence in their conclusions.
Step 3 – Use software to predict the structure
Prediction algorithms may be available as open-source software, cloud-based services, or commercial applications. Open-source implementations often require specialized hardware, command-line operation, and extensive software installation. In contrast, Lasergene Protein simplifies the process by allowing researchers to submit predictions through an intuitive graphical interface while computationally intensive processing occurs in the background.
Regardless of the algorithm selected, the goal is the same: transform a protein sequence into a three-dimensional structural model that can be explored and analyzed.
Step 4 – Analyze and interpret the results
If you want to view and analyze a structure that is already in the PDB database, you can use the built-in three-dimensional (3D) structure viewer at RSCB. But when you are dealing with a novel structure prediction, you will need to utilize a viewer that can open locally saved protein structures. As with the prediction step, protein structure viewers are available in a variety of commercial and open-source varieties (open-source application list on the PDB website).
Once a set of models has been predicted, the next steps are to evaluate prediction confidence, compare alternative models, examine protein domains and interaction interfaces, and compare predictions with experimentally determined structures, if available. Visualization tools make it possible to investigate structural features, identify potential binding sites, generate publication-quality images, and better understand how sequence influences function.
In the next chapter, we’ll explore a variety of popular modern prediction methods and provide guidance on how to choose the most suitable algorithm for your specific research needs.