bioRxiv Science⌕ Search
Home/Archive/bioRxiv

bioRxiv · 10.1101/240580

Dynamic Optimization with Particle Swarms (DOPS): A meta-heuristic for parameter estimation in biochemical models

Varner, J.·Sagar, A.·LeCover, R.·Shoemaker, C.
View original source ↗Export metadata ↓

Abstract

BackgroundMathematical modeling is a powerful tool to analyze, and ultimately design biochemical networks. However, the estimation of the parameters that appear in biochemical models is a significant challenge. Parameter estimation typically involves expensive function evaluations and noisy data, making it difficult to quickly obtain optimal solutions. Further, biochemical models often have many local extrema which further complicates parameter estimation. Toward these challenges, we developed Dynamic Optimization with Particle Swarms (DOPS), a novel hybrid meta-heuristic that combined multi-swarm particle swarm optimization with dynamically dimensioned search (DDS). DOPS uses a multi-swarm particle swarm optimization technique to generate candidate solution vectors, the best of which is then greedily updated using dynamically dimensioned search.\n\nResultsWe tested DOPS using classic optimization test functions, biochemical benchmark problems and real-world biochemical models. We performed [Formula] trials with [Formula] function evaluations per trial, and compared the performance of DOPS with other commonly used meta-heuristics such as differential evolution (DE), simulated annealing (SA) and dynamically dimensioned search (DDS). On average, DOPS outperformed other common meta-heuristics on the optimization test functions, benchmark problems and a real-world model of the human coagulation cascade.\n\nConclusionsDOPS is a promising meta-heuristic approach for the estimation of biochemical model parameters in relatively few function evaluations. DOPS source code is available for download under a MIT license at http://www.varnerlab.org.

Source connections

Explore related subjects

systems biologyPreprint

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Varner, J., Sagar, A., LeCover, R., Shoemaker, C.. 2017-12-28. Dynamic Optimization with Particle Swarms (DOPS): A meta-heuristic for parameter estimation in biochemical models. https://doi.org/10.1101/240580

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Uncovering disease progression trajectories across multiple timescales with hybrid ODE-mixture models

Biological systems are capable of evolving across multiple timescales, yet most modeling approaches capture either short-term physiological responses or long-term disease progression. Additionally, methods that combine these time scales either struggle with inter-individual variation, or require large volumes of data to work. We present Latent Class Parameter Trajectory Analysis (LCPTA), a hybrid framework that combines mechanistic ordinary differential equation (ODE) modeling of short-term dynamics with latent class trajectory modeling of long-term parameter evolution. To support this approach, we developed TemporalMixtureModels.jl, a Julia package for univariate and multivariate temporal mixture modeling. We evaluated LCPTA in two synthetic longitudinal studies, covering a simulation of a heterogeneous population developing hypertension, as well as a synthetic longitudinal study of 75 virtual individuals followed over 30 years, representing muscle insulin resistance, liver insulin resistance, and healthy trajectories. We compared LCPTA with a conditional neural ODE and a conditional neural network for reconstruction of long-term parameter trajectories and subgroup separation. In both studies LCPTA recovered relevant subgroup structure and trajectory patterns. In the cardiovascular study, the neural approaches were able to distinguish inter-individual differences, but in the insulin sensitivity study, neural approaches failed to represent parameter trajectories faithfully, primarily separating individuals along a single dominant parameter and showing weaker multivariate trajectory reconstruction. In additional experiments with sparse and irregularly sampled unseen individuals, LCPTA maintained strong classification performance. We show that LCPTA is an efficient and interpretable strategy for multiscale modeling of heterogeneous metabolic disease progression, with potential utility for data-efficient digital twin applications.

systems biology↗
bioRxiv2026

The Global Phosphorylation Landscape of Acute Heat Stress in Reef-Building Corals

Warming oceans disrupt the coral-algal symbiosis that sustains reefs, yet the signaling mechanisms initiating bleaching remain poorly understood. Integrating global phosphoproteomics across three reef-building corals with protein structure modeling revealed conservation between coral and human kinases, enabling inference of coral kinase activities and uncovering a shared signaling program induced by acute heat stress. Heat rapidly activated ERK, JNK, and p38 MAP kinases, suppressed cell cycle signaling, and remodeled pathways controlling chromatin, RNA processing, and metabolism. Most strikingly, thermal stress rewired immune receptor signaling networks associated with non-self recognition, suggesting that heat shifts the coral host from tolerance to immune recognition of its algal symbionts. These findings define conserved signaling mechanisms that precede bleaching, revealing actionable molecular targets for enhancing coral resilience.

systems biology↗
bioRxiv2026

Probing bacterial fitness in virtual chemostats

The growth rate of a bacterial population is a fundamental quantity in microbiology, but hard to predict from mechanistic single-cell models. Simulating exponentially growing cell populations is computationally infeasible, and single-lineage estimates are biased as they do not account for phenotypic selection between cells. We present computational algorithms to efficiently estimate growth rates with arbitrary accuracy by combining ideas from branching processes and particle filtering. These can be used with general agent-based cell models at a cost quadratic in the desired accuracy, bypassing the exponential complexity of naive approaches. Applying these methods to single-cell models of Escherichia coli, we study how single-cell heterogeneity in antibiotic responses shapes population dynamics, and how bacterial locomotion and chemotaxis in fluctuating environments affect their long-term growth.

systems biology↗