Coevolution Contact Map: How AI Reads Structure from…

What Coevolution Contact Map Is

Coevolution contact maps are a method used to infer the three-dimensional structure of proteins from their amino acid sequences. By analyzing how mutations in one position affect the likelihood of mutations at other positions, these maps reveal which residues are likely to be in close proximity within the protein's native state.

This technique is particularly powerful because it leverages the fact that residues that interact with each other often coevolve; changes in one residue can predict changes in another due to their functional or structural interdependence.

How Coevolution Contact Maps Work

The process begins by generating a multiple sequence alignment (MSA) of homologous proteins. Each position in the MSA is then scored based on how frequently certain pairs of residues co-occur across the aligned sequences, which is quantified using mutual information. This scoring helps to identify residue pairs that are likely to be in close contact within the protein’s structure.

Once the scores are calculated, a 3D model can be constructed by predicting the most probable contacts between these residues, effectively folding the sequence into its predicted three-dimensional form.

live demo · related simulation● LIVE

Why It Matters

Coevolution contact maps are essential for understanding protein function and structure. By accurately predicting protein structures, researchers can gain insights into how proteins fold, interact with other molecules, and perform their biological roles.

This technology is also critical in drug discovery, as it helps identify potential binding sites for drugs or other small molecules that could modulate the activity of a target protein.

Real-World Applications

Coevolution contact maps have been used to predict structures of proteins with unknown folds, which is particularly challenging. For example, they have helped in understanding the structure and function of viral proteins that are targets for antiviral drugs.

Moreover, these methods are being integrated into machine learning models like AlphaFold, which has revolutionized structural biology by providing accurate predictions without the need for experimental data.

Frequently asked questions

How does coevolution differ from direct contact in proteins?

Coevolution refers to the indirect relationship between residues that are not directly in contact but influence each other's evolution due to functional or structural dependencies, whereas direct contacts involve residues that physically interact within the protein’s native state.

What is mutual information and how does it relate to coevolution?

Mutual information measures the amount of information one random variable provides about another. In the context of coevolution, it quantifies the dependence between residue positions in a multiple sequence alignment, indicating which pairs are likely to be in close contact.

Why is AlphaFold significant in bioinformatics?

AlphaFold has significantly advanced protein structure prediction by using deep learning techniques to predict structures based on amino acid sequences alone. This has greatly reduced the need for experimental methods, which are often time-consuming and expensive.

Can coevolution contact maps be used for all types of proteins?

Coevolution contact maps work best for proteins with homologous sequences in databases. For novel or highly divergent proteins without close relatives, the accuracy can be lower, and additional experimental data may be required.

Try it live

Everything above runs in your browser — open Coevolution Contact Map: How AI Reads Structure from Sequence and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.

▶ Open Coevolution Contact Map: How AI Reads Structure from Sequence simulation