Improved hidden Markov models for molecular motors, part 2: extensions and application to experimental data

Area of Science:

• Biophysics
• Computational Biology
• Biochemistry

Background:

• Analyzing noisy single molecular motor recordings is challenging.
• Existing Markov-model based techniques have limitations in handling arbitrary step sizes and measurement noise.

Purpose of the Study:

• To extend the variable-stepsize hidden Markov model (VS-HMM) framework for improved analysis of experimental single molecular motor data.
• To develop a more robust algorithm that accounts for data-acquisition processes and random baseline drifts.

Main Methods:

• Developed the variable-stepsize integrating-detector hidden Markov model (VSI-HMM).
• Incorporated maximum a posteriori estimation as an extension.
• Utilized simulations to test the fidelity of the VSI-HMM.
• Applied the algorithm to in vitro myosin V data and in vivo melanosome motion recordings.

Main Results:

• The VSI-HMM outperforms conventional step detectors as a blind step detector.
• Identified a small population of 10 nm steps in in vitro myosin V data.
• Detected strong evidence for repeated, bidirectional steps smaller than 8 nm in in vivo melanosome motion.

Conclusions:

• The VSI-HMM provides a robust and accurate method for interpreting noisy single molecular motor recordings.
• The findings suggest that multiple motors may simultaneously carry cargo during melanosome transport.
• The VSI-HMM framework offers significant improvements for characterizing motor protein dynamics.