PhyloSDF: Phylogenetically-Conditioned Neural Generation of 3D Skull Morphology via Residual Flow Matching

2026-04-28

AI Summary

This work addresses the challenge of generating biologically plausible 3D morphologies consistent with phylogenetic relationships under extreme data scarcity in computational evolutionary biology. The authors propose a neural generative model that integrates phylogenetic constraints by embedding evolutionary relationships into the latent space. The generation process is decomposed into species-centered lookup and residual prediction, while a phylogenetic consistency loss aligns the latent space with evolutionary distances. Coupled with a residual conditional flow matching architecture, this approach enhances few-shot generation quality. Evaluated on cranial data from 24 Darwin’s finch species, the model produces 180 non-memorized meshes capturing 88–129% of real intraspecific variation and significantly outperforms diffusion models, standard flow matching, and Gaussian mixture baselines in Chamfer distance and Morphological Fréchet Distance. The method also successfully enables cross-species extrapolation and ancestral morphology reconstruction.

Abstract

Generating novel, biologically plausible three-dimensional morphological structures is a fundamental challenge in computational evolutionary biology, hampered by extreme data scarcity and the requirement that generated shapes respect phylogenetic relationships among species. In this work, we present PhyloSDF, a phylogenetically-conditioned neural generative model for 3D biological morphology that integrates two innovations: (1) a DeepSDF auto-decoder regularized by a novel Phylogenetic Consistency Loss that structures the latent space to correlate with evolutionary distances (Pearson $r=0.993$); (2) a Residual Conditional Flow Matching (Residual CFM) architecture that factorizes generation into analytic species-centroid lookup and learned residual prediction, enabling generation from as few as ~4 specimens per species. We evaluate PhyloSDF on 100 micro-CT-scanned skulls of Darwin's Finches and their relatives across 24 species. The model generates novel meshes achieving 88-129% of real intra-species variation at the code level, with all 180 generated meshes verified as non-memorized. Residual CFM surpasses denoising diffusion (which fails entirely at this scale), standard flow matching (which mode-collapses to 3-6% variation), and a Gaussian mixture baseline in both fidelity (Chamfer Distance 0.00181 vs. 0.00190) and morphometric Fréchet distance (10,641 vs. 13,322). Leave-one-species-out experiments across 18 species demonstrate phylogenetic extrapolation capability, and smooth latent interpolations produce biologically plausible ancestral skull reconstructions.

Problem

• 3D morphology generation
• phylogenetic relationships
• data scarcity
• computational evolutionary biology
• biological plausibility

Innovation

• PhyloSDF
• Phylogenetic Consistency Loss
• Residual Conditional Flow Matching
• DeepSDF
• 3D Morphological Generation