Pheno-GS: Phenoscape-scale Geodesic Sinkhorn
Abstract
High-throughput single-cell data is now collected across large patient cohorts. Understanding patient-level heterogeneity from cellular-level data motivates phenoscaping: embedding each single-cell distribution as a "datapoint," with distances given by optimal transport (OT). Computing geometry-aware OT at this scale, between all pairs of patient datasets, remains an open challenge, since existing methods either rely on Euclidean ground metrics that distort manifold structure or fail under sparse, unevenly sampled, or large-scale data. We present \textbf{Pheno-GS} (Phenoscape-scale Geodesic Sinkhorn), which computes accurate, scalable geodesic transport distances under noisy, unbalanced, large-scale settings via three components: ($1$) graph connectivity regularization for well-defined geodesics on sparse/disconnected manifolds; ($2$) an unbalanced OT formulation via KL marginal penalties; and ($3$) a batched matrix algorithm computing all pairwise distances in one heat diffusion (over 200 × faster than Geodesic Sinkhorn for $500$ distributions). We validate Pheno-GS on synthetic benchmarks and a CyTOF perturbation dataset.
Related
- Large-scale functional overlap between dorsal and ventral object-responsive networks
- Scale-dependent associations between genetic similarity and aggregation behaviour in three toad lineages
- Principles of coarse-scale functional organization in occipitotemporal cortex
- Spatial segregation and multi-scale space-use adjustments to socio-ecological variables in western gorillas
- NeuroGraphBench: Interacting with Drosophila Connectomes at Scale for Exploring the Functional Logic of Neural Circuits
BibTeX
@misc{wilkinson2026phenogs,
title = {{Pheno-GS: Phenoscape-scale Geodesic Sinkhorn}},
author = {Alistair Wilkinson and Christopher J. Tape and Smita Krishnaswamy},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2609.27633},
eprint = {2609.27633},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.27633},
}
RIS
TY - UNPB
TI - Pheno-GS: Phenoscape-scale Geodesic Sinkhorn
AU - Alistair Wilkinson
AU - Christopher J. Tape
AU - Smita Krishnaswamy
PY - 2026
DA - 2026/10/01
PB - arXiv
DO - 10.48550/arxiv.2609.27633
UR - https://arxiv.org/abs/2609.27633
AB - High-throughput single-cell data is now collected across large patient cohorts. Understanding patient-level heterogeneity from cellular-level data motivates phenoscaping: embedding each single-cell distribution as a "datapoint," with distances given by optimal transport (OT). Computing geometry-aware OT at this scale, between all pairs of patient datasets, remains an open challenge, since existing methods either rely on Euclidean ground metrics that distort manifold structure or fail under sparse, unevenly sampled, or large-scale data. We present \textbf{Pheno-GS} (Phenoscape-scale Geodesic Sinkhorn), which computes accurate, scalable geodesic transport distances under noisy, unbalanced, large-scale settings via three components: ($1$) graph connectivity regularization for well-defined geodesics on sparse/disconnected manifolds; ($2$) an unbalanced OT formulation via KL marginal penalties; and ($3$) a batched matrix algorithm computing all pairwise distances in one heat diffusion (over 200 × faster than Geodesic Sinkhorn for $500$ distributions). We validate Pheno-GS on synthetic benchmarks and a CyTOF perturbation dataset.
ER -