Protein – Structure Prediction

Introduction

The prediction of protein structure from a sequence is one of the most focused problems for researchers. This is a very useful application of bioinformatics since experimental techniques such as X-ray crystallography are very time consuming. The fundamental question is how we can predict the three-dimensional shape of a protein from its amino acid sequence. We will now see how to predict protein structure and function based on amino acid sequence.

The protein folding problem

According to Alfinsen’s hypothesis, the three-dimensional structure of a protein is determined solely by amino acid sequence information. The strong argument against Alfinsen’s hypothesis is Levinthal’s paradox. There have been some reflections on the resolution of Levinthal’s paradox. These are summarized below:

1. Theoretical methods used to test hardness are not what nature is trying to optimize.

2. Evolution may have selected for proteins that fold easily.

3. Proteins may fold locally, not globally optimally.

To summarize, it is difficult to predict protein structure from sequence. However, from the growing database of experimentally determined protein structures, some heuristics are emerging:

1. The number of unique protein folds is quite limited.

2. There are many proteins with the same fold, but no sequence similarity.

3. There are likely to be ‘neutral’ mutations that alter the structure of the protein.

Identification and characterization of proteins

Some of the tools that help predict the physical properties of known proteins are:

1. AA CompIdent.

2. TagIdent, PeptIdent and MultiIdent.

3.PROPSEARCH.

4. PepSea.

5. PepMapper, Mascot and PeptideSearch.

6. FindPept.

1. AA CompIdent:

This is used to identify proteins by their amino acid composition. It uses the amino acid composition of an unknown protein to identify known proteins of the same composition.

2. TagIdent, PeptIdent and MultiIdent:

TagIdent allows you to generate a list of proteins close to a given pI and Mw.

PeptIdent is used to identify proteins with peptide mass, pI, and Mw fingerprint data.

MultiIdent is a tool that enables protein identification using pI, Mw, amino acid composition, sequence tag, and peptide mass fingerprint data.

3.PROPSEARCH:

This is a tool to find families of putative proteins. It uses the amino acid composition as input. In addition, other properties such as molecular weight, bulky residue content, small residue content, average hydrophobicity, average charge, and content of selected dipeptide groups are also calculated from the sequence.

4. PepSea:

It is a tool for protein identification by peptide mapping or peptide sequencing.

5. PepMapper, Mascot and PeptideSearch:

PepMapper takes peptide mass as key input.

The pet search takes as input the peptide mass fingerprint, sequence query, or MS/MS ion search.

PeptideSearch uses a list of peptide masses, a peptide sequence tag, and an amino acid sequence as input.

6. FindPept:

It is used to identify peptides that result from nonspecific cleavage of proteins. This takes into account artificial chemical modifications, post-translational modifications, and autolytic protease cleavage.

Conclusion:

These are some of the protein structure and function prediction tools. This leads to the next step, which is the analysis and prediction of the primary structure.