CS Mukhopadhyay and HK Manku
School of Animal Biotechnology, GADVASU, Ludhiana
Threading or fold recognition is the method for protein tertiary structure prediction, and is implemented when no significant homologs are searched. The underlying theory assumes the existence of a limited number of distinct protein folds. The threading method thus searches the structural analogs of the query sequence in a library of representative structures through comparative fold recognition. The threading approach relies on fitting the target peptide sequence to one or more entries of a library of protein structures (e.g., protein domains, peptide chains or conserved protein cores). The best fit is evaluated on the basis of an appropriate energy function (determined by a biocomputational algorithm) to determine the best possible templates. In this way, folds are recognized simultaneously, and the whole structure can be predicted.
To predict the tertiary structure of a protein sequence using the fold recognition method.
3D structure prediction using the threading method is explained by using the RaptorX server:
The Results window displays the results as follows:
Section 1 shows the amino acid residues of the query sequence which are modeled and non‐modeled. Modeled residues are labeled as 1, whereas non‐modeled residues are labeled as 0 (shown in (A) of Figure 30.2). Modeled residues are considered as a functional part of the protein, and non‐modeled residues are considered as side loops. Moreover, the result shows the best template selection by GDT (Global Distance Test) score. For a protein with > 100 residues, GDT > 50 is a good indicator (shown in (B) of Figure 30.2). In this example protein sequence, accession no. P79362 in SwissProt has been taken for tertiary structure prediction.
This alignment (shown in (C) of Figure 30.2) is used for constructing the 3D model. The chemical nature of the residues in the alignment dictates the colors of each position:
The predicted results indicated that one domain that had the best matching template; 1n5hA had the P‐value of 6.78e–09. The P‐value is an indicator of the goodness of the best template representing the predicted secondary structure. Overall, 64% of the residues were modeled.
The user can switch between the pairs of models (three state and eight state models), or use both of them for secondary structure prediction. Hovering over a residue will display the exact distribution of secondary structure classes in a pop‐up box appearing next to the residue. The three‐state secondary structure types are also abbreviated as “H”, “E”, and “C”, which represent helix, beta sheet, and loop, respectively.
This results window shows the three‐class secondary structure result. A single block contains ten residues, which together form a histogram‐like shape in a block, where each bar indicates the percentage of helix, coil, and beta strand. For example, P at the 135th position contributes 15.5% helix, 18.8% beta strand, and 67.8% coil (as shown in (A) in Figure 30.4). Similarly, an eight‐class secondary structure result has been demonstrated, showing the percentage of alpha helix, isolated beta bridge, extended strand in the beta ladder, 3‐helix, 5‐helix, hydrogen bonded turn, bend, and coil (as shown in (B) in Figure 30.4).
In addition to the structural element prediction, RaptorX also predicts the percentages of the residues which contribute to the disordered conformational randomness of the secondary structure. Moreover, it gives the percentages of the residues which are involved in the formation of the solvent‐accessible surface. Maximum exposure of the residue means a greater contribution towards the solvent‐accessible surface. For example, I at the 122nd position is 81.8% buried, 15.4% in the medium region, and 2.7% exposed. This means that this residue makes the smallest contribution towards the formation of the solvent‐accessible surface.
The threading method is beneficial for the identification of the conserved domains. The predicted tertiary structure can be further analyzed for model quality validation using Procheck.
3.23.101.203