Limits...
Improving protein-ligand binding site prediction accuracy by classification of inner pocket points using local features.

Krivák R, Hoksza D - J Cheminform (2015)

Bottom Line: The ligandability scores are finally merged into the resulting pocket score to be used for prioritization of the putative pockets.With the used of multiple datasets the experimental results demonstrate that the application of our method as a post-processing step greatly increases the quality of the prediction of Fpocket and ConCavity, two state of the art protein-ligand binding site prediction algorithms.The positive experimental results show that our method can be used to improve the success rate, validity and applicability of existing protein-ligand binding site prediction tools.

View Article: PubMed Central - PubMed

Affiliation: Department of Software Engineering, Charles University in Prague, Prague, Czech Republic.

ABSTRACT

Background: Protein-ligand binding site prediction from a 3D protein structure plays a pivotal role in rational drug design and can be helpful in drug side-effects prediction or elucidation of protein function. Embedded within the binding site detection problem is the problem of pocket ranking - how to score and sort candidate pockets so that the best scored predictions correspond to true ligand binding sites. Although there exist multiple pocket detection algorithms, they mostly employ a fairly simple ranking function leading to sub-optimal prediction results.

Results: We have developed a new pocket scoring approach (named PRANK) that prioritizes putative pockets according to their probability to bind a ligand. The method first carefully selects pocket points and labels them by physico-chemical characteristics of their local neighborhood. Random Forests classifier is subsequently applied to assign a ligandability score to each of the selected pocket point. The ligandability scores are finally merged into the resulting pocket score to be used for prioritization of the putative pockets. With the used of multiple datasets the experimental results demonstrate that the application of our method as a post-processing step greatly increases the quality of the prediction of Fpocket and ConCavity, two state of the art protein-ligand binding site prediction algorithms.

Conclusions: The positive experimental results show that our method can be used to improve the success rate, validity and applicability of existing protein-ligand binding site prediction tools. The method was implemented as a stand-alone program that currently contains support for Fpocket and Concavity out of the box, but is easily extendible to support other tools. PRANK is made freely available at http://siret.ms.mff.cuni.cz/prank.

No MeSH data available.


Related in: MedlinePlus

Visualization of inner pocket points.(a) Displayed is protein 1AZM from DT198 dataset bound to one ligand (magenta). Fpocket predicted 13 pockets that are depicted as colored areas on the protein surface. To rank these pockets, the protein was first covered with evenly spaced Connolly surface points (probe radius 1.6 Å) and only the points adjacent to one of the pockets were retained. Color of the points reflects their ligandability (green = 0…red = 0.7) predicted by Random Forest classifier. PRANK algorithm rescores pockets according to the cumulative ligandability of their corresponding points. Note that there are two clusters of ligandable points in the picture, one located in the upper dark-blue pocket and the other in the light-blue pocket in the middle. The light-blue pocket, which is in fact the true binding site, contains more ligandable points and therefore will be ranked higher. (b) Detailed view of the binding site with ligand and inner pocket points.
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC4414931&req=5

Fig2: Visualization of inner pocket points.(a) Displayed is protein 1AZM from DT198 dataset bound to one ligand (magenta). Fpocket predicted 13 pockets that are depicted as colored areas on the protein surface. To rank these pockets, the protein was first covered with evenly spaced Connolly surface points (probe radius 1.6 Å) and only the points adjacent to one of the pockets were retained. Color of the points reflects their ligandability (green = 0…red = 0.7) predicted by Random Forest classifier. PRANK algorithm rescores pockets according to the cumulative ligandability of their corresponding points. Note that there are two clusters of ligandable points in the picture, one located in the upper dark-blue pocket and the other in the light-blue pocket in the middle. The light-blue pocket, which is in fact the true binding site, contains more ligandable points and therefore will be ranked higher. (b) Detailed view of the binding site with ligand and inner pocket points.

Mentions: Individual steps are described in greater detail in following sections. For the visualization of classified pocket points see Figure 2.Figure 2


Improving protein-ligand binding site prediction accuracy by classification of inner pocket points using local features.

Krivák R, Hoksza D - J Cheminform (2015)

Visualization of inner pocket points.(a) Displayed is protein 1AZM from DT198 dataset bound to one ligand (magenta). Fpocket predicted 13 pockets that are depicted as colored areas on the protein surface. To rank these pockets, the protein was first covered with evenly spaced Connolly surface points (probe radius 1.6 Å) and only the points adjacent to one of the pockets were retained. Color of the points reflects their ligandability (green = 0…red = 0.7) predicted by Random Forest classifier. PRANK algorithm rescores pockets according to the cumulative ligandability of their corresponding points. Note that there are two clusters of ligandable points in the picture, one located in the upper dark-blue pocket and the other in the light-blue pocket in the middle. The light-blue pocket, which is in fact the true binding site, contains more ligandable points and therefore will be ranked higher. (b) Detailed view of the binding site with ligand and inner pocket points.
© Copyright Policy - open-access
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC4414931&req=5

Fig2: Visualization of inner pocket points.(a) Displayed is protein 1AZM from DT198 dataset bound to one ligand (magenta). Fpocket predicted 13 pockets that are depicted as colored areas on the protein surface. To rank these pockets, the protein was first covered with evenly spaced Connolly surface points (probe radius 1.6 Å) and only the points adjacent to one of the pockets were retained. Color of the points reflects their ligandability (green = 0…red = 0.7) predicted by Random Forest classifier. PRANK algorithm rescores pockets according to the cumulative ligandability of their corresponding points. Note that there are two clusters of ligandable points in the picture, one located in the upper dark-blue pocket and the other in the light-blue pocket in the middle. The light-blue pocket, which is in fact the true binding site, contains more ligandable points and therefore will be ranked higher. (b) Detailed view of the binding site with ligand and inner pocket points.
Mentions: Individual steps are described in greater detail in following sections. For the visualization of classified pocket points see Figure 2.Figure 2

Bottom Line: The ligandability scores are finally merged into the resulting pocket score to be used for prioritization of the putative pockets.With the used of multiple datasets the experimental results demonstrate that the application of our method as a post-processing step greatly increases the quality of the prediction of Fpocket and ConCavity, two state of the art protein-ligand binding site prediction algorithms.The positive experimental results show that our method can be used to improve the success rate, validity and applicability of existing protein-ligand binding site prediction tools.

View Article: PubMed Central - PubMed

Affiliation: Department of Software Engineering, Charles University in Prague, Prague, Czech Republic.

ABSTRACT

Background: Protein-ligand binding site prediction from a 3D protein structure plays a pivotal role in rational drug design and can be helpful in drug side-effects prediction or elucidation of protein function. Embedded within the binding site detection problem is the problem of pocket ranking - how to score and sort candidate pockets so that the best scored predictions correspond to true ligand binding sites. Although there exist multiple pocket detection algorithms, they mostly employ a fairly simple ranking function leading to sub-optimal prediction results.

Results: We have developed a new pocket scoring approach (named PRANK) that prioritizes putative pockets according to their probability to bind a ligand. The method first carefully selects pocket points and labels them by physico-chemical characteristics of their local neighborhood. Random Forests classifier is subsequently applied to assign a ligandability score to each of the selected pocket point. The ligandability scores are finally merged into the resulting pocket score to be used for prioritization of the putative pockets. With the used of multiple datasets the experimental results demonstrate that the application of our method as a post-processing step greatly increases the quality of the prediction of Fpocket and ConCavity, two state of the art protein-ligand binding site prediction algorithms.

Conclusions: The positive experimental results show that our method can be used to improve the success rate, validity and applicability of existing protein-ligand binding site prediction tools. The method was implemented as a stand-alone program that currently contains support for Fpocket and Concavity out of the box, but is easily extendible to support other tools. PRANK is made freely available at http://siret.ms.mff.cuni.cz/prank.

No MeSH data available.


Related in: MedlinePlus