Overlay databank unlocks data-driven analyses of biomolecules for all

Area of Science:

• Computational biology
• Biophysics
• Data science

Background:

• Artificial intelligence (AI) applications are expanding, but often lack accessible training data.
• Cellular membranes are crucial for biological functions and disease, yet challenging to study experimentally.
• Molecular Dynamics (MD) simulations offer insights into membranes but face computational and data accessibility challenges.

Purpose of the Study:

• To introduce an overlay databank format for organizing and accessing scattered data.
• To create a community-driven database (NMRlipids Databank) for cellular membrane simulations.
• To facilitate data-driven and machine learning applications in membrane biophysics.

Main Methods:

• Development of the overlay databank format for data accessibility.
• Creation of the NMRlipids Databank with quality-evaluated, atom-resolution MD simulations.
• Implementation of programmatic access and a graphical user interface (GUI) for data retrieval.

Main Results:

• Demonstrated the practical application of the overlay databank format with the NMRlipids Databank.
• Provided programmatic and GUI access to extensive MD simulation data of cellular membranes.
• Enabled enhanced analysis and understanding of membrane systems beyond traditional methods.

Conclusions:

• The overlay databank format effectively addresses data accessibility barriers for AI and machine learning.
• The NMRlipids Databank significantly advances the study of cellular membranes using simulation data.
• The proposed concept has broad applicability to other biomolecules and scientific fields facing similar data challenges.