Dataset Information

PLAS-20k: Extended Dataset of Protein-Ligand Affinities from MD Simulations for Machine Learning Applications.

ABSTRACT: Computing binding affinities is of great importance in drug discovery pipeline and its prediction using advanced machine learning methods still remains a major challenge as the existing datasets and models do not consider the dynamic features of protein-ligand interactions. To this end, we have developed PLAS-20k dataset, an extension of previously developed PLAS-5k, with 97,500 independent simulations on a total of 19,500 different protein-ligand complexes. Our results show good correlation with the available experimental values, performing better than docking scores. This holds true even for a subset of ligands that follows Lipinski's rule, and for diverse clusters of complex structures, thereby highlighting the importance of PLAS-20k dataset in developing new ML models. Along with this, our dataset is also beneficial in classifying strong and weak binders compared to docking. Further, OnionNet model has been retrained on PLAS-20k dataset and is provided as a baseline for the prediction of binding affinities. We believe that large-scale MD-based datasets along with trajectories will form new synergy, paving the way for accelerating drug discovery.

SUBMITTER: Korlepara DB

PROVIDER: S-EPMC10858175 | biostudies-literature | 2024 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

PLAS-20k: Extended Dataset of Protein-Ligand Affinities from MD Simulations for Machine Learning Applications.

Korlepara Divya B DB C S Vasavi V Srivastava Rakesh R Pal Pradeep Kumar PK Raza Saalim H SH Kumar Vishal V Pandit Shivam S Nair Aathira G AG Pandey Sanjana S Sharma Shubham S Jeurkar Shruti S Thakran Kavita K Jaglan Reena R Verma Shivangi S Ramachandran Indhu I Chatterjee Prathit P Nayar Divya D Priyakumar U Deva UD

Scientific data 20240209 1

Computing binding affinities is of great importance in drug discovery pipeline and its prediction using advanced machine learning methods still remains a major challenge as the existing datasets and models do not consider the dynamic features of protein-ligand interactions. To this end, we have developed PLAS-20k dataset, an extension of previously developed PLAS-5k, with 97,500 independent simulations on a total of 19,500 different protein-ligand complexes. Our results show good correlation wit ...[more]

PMID: 38336857

Similar Datasets

Project description:BACKGROUND:The role of dynamics in protein functions including signal transduction is just starting to be deciphered. Eph receptors with 16 members divided into A- and B- subclasses are respectively activated by 9 A- and B-ephrin ligands. EphA4 is the only receptor capable of binding to all 9 ephrins and small molecules with overlapped interfaces. RESULTS:We first determined the structures of the EphA4 ligand binding domain (LBD) in two crystals of P1 space group. Noticeably, 8 EphA4 molecules were found in one asymmetric unit and consequently from two crystals we obtained 16 structures, which show significant conformational variations over the functionally critical A-C, D-E, G-H and J-K loops. The 16 new structures, together with previous 9 ones, can be categorized into two groups: closed and open forms which resemble the uncomplexed and complexed structures of the EphA4 LBD respectively. To assess whether the conformational diversity over the loops primarily results from the intrinsic dynamics, we initiated 30-ns molecular dynamics (MD) simulations for both closed and open forms. The results indicate that the loops do have much higher intrinsic dynamics, which is further unravelled by NMR H/D exchange experiments. During simulations, the open form has the RMS deviations slightly larger than those of the closed one, suggesting the open form may be less stable in the absence of external contacts. Furthermore, no obvious exchange between two forms is observed within 30 ns, implying that they are dynamically separated. CONCLUSIONS:Our study provides the first experimental and computational result revealing that the intrinsic dynamics are most likely underlying the conformational diversity observed for the EphA4 LBD loops mediating the binding affinity and specificity. Interestingly, the open conformation of the EphA4 LBD is slightly unstable in the absence of it natural ligand ephrins, implying that the conformational transition from the closed to open has to be driven by the high-affinity interaction with ephrins because the weak interaction with small molecule was found to be insufficient to trigger the transition. Our results therefore highlight the key role of protein dynamics in Eph-ephrin signalling and would benefit future design of agonists/antagonists targeting Eph receptors.

Dataset Information

PLAS-20k: Extended Dataset of Protein-Ligand Affinities from MD Simulations for Machine Learning Applications.

Publications

PLAS-20k: Extended Dataset of Protein-Ligand Affinities from MD Simulations for Machine Learning Applications.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets