[Paper Review] An extension to reversible jump Markov chain Monte Carlo for change point problems with heterogeneous temporal dynamics

Emily M. Gribbin, Benjamin Davis+2 more|arXiv (Cornell University)|Feb 19, 2026

Advanced Fluorescence Microscopy Techniques0 citations

TL;DR

The paper extends RJMCMC with compound change point moves (CRJMCMC) to detect short-lived state changes in time series, applied to FLImP photobleach step analysis, improving per-frame fluorophore counting without calibration.

ABSTRACT

Detecting brief changes in time-series data remains a major challenge in fields where short-lived states carry meaning. In single-molecule localisation microscopy, this problem is particularly acute as fluorescent molecules used to tag protein oligomers display heterogenous photophysical behaviour that can complicate photobleach step analysis; a key step in resolving nanoscale protein organisation. Existing methods often require extensive filtering or prior calibration, and can fail to accurately account for blinking or reversible dark states that may contaminate downstream analysis. In this paper, an extension to RJMCMC is proposed for change point detection with heterogeneous temporal dynamics. This approach is applied to the problem of estimating per-frame active fluorophore counts from one-dimensional integrated intensity traces derived from Fluorescence Localisation Imaging with Photobleaching (FLImP), where compound change point pair moves are introduced to better account for short-lived events known as blinking and dark states. The approach is validated using simulated and experimental data, demonstrating improved accuracy and robustness when compared with current photobleach step analysis methods and with the existing analysis approach for FLImP data. This Compound RJMCMC (CRJMCMC) algorithm performs reliably across a wide range of fluorophore counts and signal-to-noise conditions, with signal-to-noise ratio (SNR) down to 0.001 and counts as high as nineteen fluorophores, while also effectively estimating low counts observed when studying EGFR oligomerisation. Beyond single molecule imaging, this work has applications for a variety of time series change point detection problems with heterogeneous state persistence. For example, electrocorticography brain-state segmentation, fault detection in industrial process monitoring and realised volatility in financial time series.

Motivation & Objective

• Address the challenge of detecting brief, meaningful state changes in time-series with heterogeneous dynamics.
• Develop a scalable RJMCMC extension that jointly adds/removes close change points to model short-lived events.
• Apply the method to one-dimensional integrated intensity traces from FLImP to estimate per-frame active fluorophore counts.
• Demonstrate reduced reliance on prior calibration and user input in photobleach step analysis.

Proposed method

• Introduce a multiple change point model with k change points and Poisson-distributed k.
• Use Gaussian noise for frame intensities with mean mu_i = mu_f n_i + mu_b and var sigma_i^2 = sigma_f^2 n_i + sigma_b^2, where n_i are active fluorophores.
• Propose compound birth-death moves that add or remove pairs of closely spaced change points to capture short-lived blink/dark states.
• Maintain RJMCMC structure with birth, death, and shift moves, but custom-tailor moves to accommodate integer active-fluorophore counts.
• Initialize and update intensity parameters mu_f, mu_b, sigma_f^2, sigma_b^2 with priors to avoid heavy calibration.
• Validate against simulated and experimental FLImP traces, comparing RMSE to state-of-the-art methods and FLImP’s existing analysis.

Experimental results

Research questions

• RQ1 Can CRJMCMC reliably detect short-lived blink and dark states in photobleach traces without prior calibration?
• RQ2 How does CRJMCMC perform in estimating per-frame active fluorophore counts across varying fluorophore numbers, SNR, and state-transition frequencies?
• RQ3 How does the CRJMCMC approach compare to MAP-based and HMM-MCMC methods in terms of accuracy and robustness for FLImP data?

Key findings

• CRJMCMC maintains low RMSE up to seventeen fluorophores and remains robust as SNR decreases to 0.001.
• CRJMCMC outperforms monotonic MAP and sequential MAP, and is competitive with factorial HMM-MCMC, especially at low fluorophore counts.
• CRJMCMC handles short-lived states without heavy filtering or labelled calibration, increasing usable frames by about 30% on FLImP DNA origami data.
• Average run time for CRJMCMC is 89.31 s per trace, faster than factorial HMM-MCMC (324.81 s) and slower than MAP methods, with CPU parallelism suggesting further speed gains.
• On experimental FLImP traces, CRJMCMC achieves high frame-wise accuracy, precision, and sensitivity across two-, three-, and four-fluorophore levels when compared to FLImP ground-truth tracks.