Bound2Learn: a machine learning approach for classification of DNA-bound proteins from single-molecule tracking...

Area of Science:

• Molecular Biology
• Genomics
• Biophysics

Background:

• DNA-bound proteins are crucial for genome maintenance and regulation.
• Understanding protein-DNA interaction kinetics, such as residence times, offers insights into biological processes.
• Direct visualization of protein-DNA kinetics via single-particle tracking is challenging due to difficulties in identifying bound states and slow kinetics.

Purpose of the Study:

• To develop a robust machine learning approach for classifying protein tracks and accurately estimating DNA-bound protein residence times.
• To overcome limitations in current single-particle tracking analysis for slow kinetic processes.

Main Methods:

• Developed Bound2Learn, a machine learning algorithm.
• Utilized output from established tracking software for track classification.
• Validated the approach using in silico simulations and live-cell data.

Main Results:

• Bound2Learn accurately classifies protein tracks, enabling precise estimation of residence times.
• The method was successfully validated in both simulated environments and live-cell experiments.
• Demonstrated efficacy in Escherichia coli and Saccharomyces cerevisiae.

Conclusions:

• Bound2Learn provides a robust solution for analyzing protein-DNA kinetics.
• The method enhances the study of genome regulation by accurately quantifying protein residence times.
• This approach has broad applicability across different organisms and biological systems.