Machine learning applied to estimate the impact of mutations at protein-protein interfaces based on physico-chemical, statistical and evolutionary conservation descriptors
Trabajo presentado en el Annual General Meeting ELIXIR 3D BioInfo Community F2F (hybrid meeting), celebrado en Hinxton (Reino Unido), del 2 al 4 de noviembre de 2022
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | comunicación de congreso biblioteca |
| Published: |
2022-11-02
|
| Online Access: | http://hdl.handle.net/10261/303884 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| id |
dig-icvv-es-10261-303884 |
|---|---|
| record_format |
koha |
| spelling |
dig-icvv-es-10261-3038842023-03-22T13:00:05Z Machine learning applied to estimate the impact of mutations at protein-protein interfaces based on physico-chemical, statistical and evolutionary conservation descriptors Glaser, Fabian Rodríguez-Lumbreras, Luis A. Fernández-Recio, Juan Trabajo presentado en el Annual General Meeting ELIXIR 3D BioInfo Community F2F (hybrid meeting), celebrado en Hinxton (Reino Unido), del 2 al 4 de noviembre de 2022 Inspired by our participation in the Activity II of the ELIXIR 3D-BioInfo Community "Open resources for sharing, integrating and benchmarking software tools for modelling the proteome in 3D"., in which we applied a variety of scoring parameters from pyDock [1], CCharPPI [2] and ConSurf [3] to improve the discrimination between physiological and non-physiological dimer interfaces, we have further explored the use of these functions to estimate the impact of mutations on protein-protein binding affinity. Sets of mutants were obtained from SKEMPI v2.0 database [4], as well as their binding affinity and kinetic values and the wild type (WT) complex structures. The structure of mutants can be modelled by SCWRL [5] and by molecular dynamics (MD). Then, different descriptors were applied to the mutated and WT structures (as well as on different MD-based conformations), first individually, and then in combination obtained from random forest classifiers. Energy-based descriptors such as electrostatics, desolvation and van der Waals provided predictive results comparable to other multiparametric methods. The application of random forest classifiers to the entire set of descriptors shows promising predictive results for the detection of beneficial and deleterious mutations regarding their impact on protein-protein binding affinity. The results of this predictor are comparable with the gold-standard MM-PBSA [6] at much lower computational cost. Finally, the approach was extended to mutations affecting protein-protein interactions for which no complex structure is available, by using docking models. 2023-03-22T13:00:04Z 2023-03-22T13:00:04Z 2022-11-02 2023-03-22T13:00:04Z comunicación de congreso Annual General Meeting ELIXIR 3D BioInfo Community F2F (2022) http://hdl.handle.net/10261/303884 Sí open |
| institution |
ICVV ES |
| collection |
DSpace |
| country |
España |
| countrycode |
ES |
| component |
Bibliográfico |
| access |
En linea |
| databasecode |
dig-icvv-es |
| tag |
biblioteca |
| region |
Europa del Sur |
| libraryname |
Biblioteca del ICVV España |
| description |
Trabajo presentado en el Annual General Meeting ELIXIR 3D BioInfo Community F2F (hybrid meeting), celebrado en Hinxton (Reino Unido), del 2 al 4 de noviembre de 2022 |
| format |
comunicación de congreso |
| author |
Glaser, Fabian Rodríguez-Lumbreras, Luis A. Fernández-Recio, Juan |
| spellingShingle |
Glaser, Fabian Rodríguez-Lumbreras, Luis A. Fernández-Recio, Juan Machine learning applied to estimate the impact of mutations at protein-protein interfaces based on physico-chemical, statistical and evolutionary conservation descriptors |
| author_facet |
Glaser, Fabian Rodríguez-Lumbreras, Luis A. Fernández-Recio, Juan |
| author_sort |
Glaser, Fabian |
| title |
Machine learning applied to estimate the impact of mutations at protein-protein interfaces based on physico-chemical, statistical and evolutionary conservation descriptors |
| title_short |
Machine learning applied to estimate the impact of mutations at protein-protein interfaces based on physico-chemical, statistical and evolutionary conservation descriptors |
| title_full |
Machine learning applied to estimate the impact of mutations at protein-protein interfaces based on physico-chemical, statistical and evolutionary conservation descriptors |
| title_fullStr |
Machine learning applied to estimate the impact of mutations at protein-protein interfaces based on physico-chemical, statistical and evolutionary conservation descriptors |
| title_full_unstemmed |
Machine learning applied to estimate the impact of mutations at protein-protein interfaces based on physico-chemical, statistical and evolutionary conservation descriptors |
| title_sort |
machine learning applied to estimate the impact of mutations at protein-protein interfaces based on physico-chemical, statistical and evolutionary conservation descriptors |
| publishDate |
2022-11-02 |
| url |
http://hdl.handle.net/10261/303884 |
| work_keys_str_mv |
AT glaserfabian machinelearningappliedtoestimatetheimpactofmutationsatproteinproteininterfacesbasedonphysicochemicalstatisticalandevolutionaryconservationdescriptors AT rodriguezlumbrerasluisa machinelearningappliedtoestimatetheimpactofmutationsatproteinproteininterfacesbasedonphysicochemicalstatisticalandevolutionaryconservationdescriptors AT fernandezreciojuan machinelearningappliedtoestimatetheimpactofmutationsatproteinproteininterfacesbasedonphysicochemicalstatisticalandevolutionaryconservationdescriptors |
| _version_ |
1777671066779385856 |