optimal transport

This paper uses global optimization to infer scaling laws of branching networks from their geometry, revealing type-specific signatures and outperforming existing laws in vascular systems.

The paper presents an optimized geospatial prioritization scheme for bird conservation in Eastern Canada using linear programming to maximize population coverage.

This paper introduces a formal framework for deriving synteny blocks directly from sequence data to better capture genomic rearrangements without obscuring structural variation.

The paper develops BrainMapping to align rat and human cortices using resting-state fMRI, revealing distributed and directionally asymmetric correspondences between species.

This paper introduces UOT-IOT, an inverse optimal transport framework for single-cell state transitions, showing improved identifiability and performance over existing methods.

CardamomOT improves gene regulatory network inference and trajectory reconstruction by integrating mechanistic models with optimal transport, using exact time labels and reducing hyperparameters.

Species-OT uses Gromov-Wasserstein optimal transport to compare transcriptome geometries across species, identifying gene correspondences and transcriptomic discrepancies.

The paper optimizes multiple circular sequence alignment for cyclic peptides, achieving faster performance while maintaining alignment quality.

This paper optimises microphone array geometry for improved localisation accuracy in bioacoustic applications, showing that geometry, not sensor count, primarily determines performance.

State, a set-based transformer trained on over 100 million perturbed cells, predicts perturbation responses in new cellular contexts better than mean, linear and deep learning baselines.

The paper introduces ID3, a framework for optimizing mRNA sequences using gradient-based methods, balancing multiple properties while preserving amino acid sequences.

This study optimises fertilisation kinetics models for two Acropora coral species using experimental data to improve predictions of reproductive success and inform reef management.

The paper introduces arteMIS, a framework for optimizing molecular network parameters in metabolomics, showing improved network quality over default settings.

This paper compares Adam and Muon optimizers for training regulatory DNA Transformer models, finding Muon reduces compute costs while achieving similar performance.

UQ-PhysiCell provides a Python framework for uncertainty quantification, calibration, and model analysis in PhysiCell simulations, enabling scalable and reproducible uncertainty-aware analysis of agent-based models.

This paper introduces a method to analyze neural avalanches using optimal transport and dynamic time warping, revealing 12 propagation patterns and a shift toward stable patterns under psilocybin.

This paper introduces PGLD, an AI method that designs cost-effective DNA libraries for high-throughput experiments by optimizing synthesis-aware stochastic libraries.

The paper introduces METEOR-K, a framework that incorporates metabolite turnover rates as constraints to improve the physiological consistency of steady-state metabolic models.

ParTIpy scales archetypal analysis for biological trade-offs and integrates with single-cell workflows, enabling task inference in large datasets.

The paper evaluates deconvolution methods' robustness to incomplete reference matrices and introduces DICEPro to improve performance by adjusting reference signatures.

SIMOFF infers metabolic objectives from flux data, improving gene essentiality predictions in yeast and CHO cells.

This paper compares functional alignment methods and template construction strategies in fMRI to improve group-level decoding accuracy and interpretability.

DivQuant estimates species richness and entropy from small samples, achieving accurate confidence intervals and outperforming existing methods across multiple datasets.

This paper introduces a bi-level diversity-optimization framework for selecting representative protein panels, reducing redundancy and improving diversity compared to existing methods.

The paper introduces SPECTRA, a method to reconstruct metabolic networks from multi-omics data at large-scale, improving speed and reaction inclusion across cancer cell lines, microbial reconstructions, and synthetic gut microbiota.

This paper presents an AI framework to prioritize conservation and restoration actions that align biodiversity, climate, and water targets by optimizing habitat connectivity and co-benefits.

TOAST improves spatial omics data alignment by integrating spatial constraints into optimal transport, enhancing accuracy and biological interpretability.

The paper explains why phylogenetic compression works well by introducing a formal framework and showing that Neighbor Joining achieves near-optimal compression under the Infinite Sites Model.

FASTERCC improves metabolic network analysis by accelerating flux consistency testing and reconstruction, reducing runtime by 20-fold for large networks.

This paper introduces a method to match branch lengths between trees, improving accuracy in phylogenomic analyses by addressing the challenges of heterogeneous branch length measurements.