Attention:The NSF Public Access Repository (PAR) system and access will be unavailable from 5:00 PM ET until 8:00 PM ET on Friday, September 11 due to maintenance. We apologize for the inconvenience.


This content will become publicly available on April 21, 2027

Title: Deep learning framework for quantifying self-organization in Myxococcus xanthus
Under starvation,Myxococcus xanthusbacteria initiate a multicellular developmental program in which cells move to form fruiting bodies and differentiate into distinct cell types. Many genes affecting this process have been identified, and it is assumed that perturbing genes within the same pathway induces similar changes in the phenotype, although these changes may be subtle or obscured by pleiotropy. However, these pathways cannot be systematically mapped because there are no reliable methods for quantifying phenotype similarity. Here, we applied deep learning to quantify phenotype patterns and self-organization dynamics of 292 genetically distinct strains. We integrated ResNet and StyleGAN2 into a Variational Autoencoder and trained it together with a Siamese network that learns phenotypic similarity. This end-to-end system encoded high-resolution microscopy images into 13-dimensional feature vectors, effectively capturing variation in aggregation patterns across time and strains. Human evaluation confirmed that our model’s reconstructions were visually indistinguishable from real images and closely aligned with input phenotypes. Importantly, the feature space is interpretable: Individual dimensions correlate with biological features such as aggregate number and size, and extrapolation along these dimensions produces predictable morphological changes. Remarkably, our model revealed that developmental phenotypes and ultimate aggregation fate are predictable from the earliest images before visible aggregation begins. This predictability held across both genetic and environmental sources of variation, suggesting that subtle, early-stage phenotypic signatures carry critical information about developmental trajectories. These results demonstrate how machine learning can reveal hidden aspects of complex multicellular dynamics and provide methods for phenotypic analysis without manual annotation.  more » « less
Award ID(s):
2019745
PAR ID:
10687597
Author(s) / Creator(s):
; ; ; ; ; ; ; ;
Publisher / Repository:
PNAS
Date Published:
Journal Name:
Proceedings of the National Academy of Sciences
Volume:
123
Issue:
16
ISSN:
0027-8424
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Andrews, B (Ed.)
    Abstract Neurospora crassa propagates through dissemination of conidia, which develop through specialized structures called conidiophores. Recent work has identified striking variation in conidiophore morphology, using a wild population collection from Louisiana, United States of America to classify 3 distinct phenotypes: Wild-Type, Wrap, and Bulky. Little is known about the impact of these phenotypes on sporulation or germination later in the N. crassa life cycle, or about the genetic variation that underlies them. In this study, we show that conidiophore morphology likely affects colonization capacity of wild N. crassa isolates through both sporulation distance and germination on different carbon sources. We generated and crossed homokaryotic strains belonging to each phenotypic group to more robustly fit a model for and estimate heritability of the complex trait, conidiophore architecture. Our fitted model suggests at least 3 genes and 2 epistatic interactions contribute to conidiophore phenotype, which has an estimated heritability of 0.47. To uncover genes contributing to these phenotypes, we performed RNA-sequencing on mycelia and conidiophores of strains representing each of the 3 phenotypes. Our results show that the Bulky strain had a distinct transcriptional profile from that of Wild-Type and Wrap, exhibiting differential expression patterns in clock-controlled genes (ccgs), the conidiation-specific gene con-6, and genes implicated in metabolism and communication. Combined, these results present novel ecological impacts of and differential gene expression underlying natural conidiophore morphological variation, a complex trait that has not yet been thoroughly explored. 
    more » « less
  2. The genotype-to-phenotype problem (G2P) for multicellular development asks how genetic inputs control collective phenotypic outputs. However, this is a challenging problem due to gene redundancy and stochasticity, causing mutations to have subtle phenotypic effects and replicates to display significant variation. We approach this problem using the model organism Myxococcus xanthus, a motile self-organizing bacterium that forms three-dimensional cell aggregates that mature into spore-filled fruiting bodies when under starvation stress. We develop a high-throughput imaging method using three-dimensional-printed microscopes to efficiently collect large phenotypic datasets. Our automated methods for analysis and visualization produce a map of phenotypic variation in M. xanthus development. We demonstrate that even subtle effects on developmental dynamics caused by mutation can be identified, discriminated, characterized, and given statistical significance, with implications for future gene annotation studies and the effect of environmental factors on G2P. 
    more » « less
  3. Neurospora crassa propagates through dissemination of conidia, which develop through specialized structures called conidiophores. Recent work has identified striking variation in conidiophore morphology, using a wild population collection from Louisiana, United States of America (USA) to classify three distinct phenotypes: Wild-Type (WT), Wrap, and Bulky. Little is known about the impact of these phenotypes on sporulation or germination later in the N. crassa life cycle, or about the genetic variation that underlies them. In this study, we show that conidiophore morphology likely affects colonization capacity of wild N. crassa isolates through both sporulation distance and germination on different carbon sources. We generated and crossed homokaryotic strains belonging to each phenotypic group to more robustly fit a model for and estimate heritability of the complex trait, conidiophore architecture. Our fitted model suggests at least three genes and two epistatic interactions contribute to conidiophore phenotype, which has an estimated heritability of 0.47. To uncover genes contributing to these phenotypes, we performed RNA-Sequencing (RNA-Seq) on mycelia and conidiophores of strains representing each of the three phenotypes. Our results show that the Bulky strain had a distinct transcriptional profile from that of WT and Wrap, exhibiting differential expression patterns in clock-controlled genes (ccgs), the conidiation-specific gene con-6, and genes implicated in metabolism and communication. Combined, these results present novel ecological impacts of and differential gene expression underlying natural conidiophore morphological variation, a complex trait that has not yet been thoroughly explored. 
    more » « less
  4. Abstract Although similar developmental regulatory networks can produce diverse phenotypes, different networks can also produce the same phenotype. In theory, as long as development can produce an acceptable end phenotype, the details of the process could be shielded from selection, leading to the possibility of developmental system drift, where the developmental mechanisms underlying a stable phenotype continue to evolve. Many examples exist of divergent developmental genetics underlying conserved traits. However, studies that elucidate how these differences arose and how other features of development accommodated them are rarer. InCaenorhabditis elegans, six GATA-type transcription factors (GATA factors) comprise the zygotic part of the endoderm specification network. Here we show that the core of this network - five of the genes - originated within the genus during a brief but explosive radiation of this gene family and that at least three of them evolved from a single ancestral gene with at least two different spatio-temporal expression patterns. Based on analyses of their evolutionary history, gene structure, expression, and sequence, we explain how these GATA factors were integrated into this network. Our results show how gene duplication fueled the developmental system drift of the endoderm network in a phylogenetically brief period in developmentally canalized worms. 
    more » « less
  5. ABSTRACT Single mutations frequently alter several aspects of cell behavior but rarely reveal whether a particular statistically significant change is biologically significant. To determine which behavioral changes are most important for multicellular self-organization, we devised a new methodology using Myxococcus xanthus as a model system. During development, myxobacteria coordinate their movement to aggregate into spore-filled fruiting bodies. We investigate how aggregation is restored in two mutants, csgA and pilC , that cannot aggregate unless mixed with wild-type (WT) cells. To this end, we use cell tracking to follow the movement of fluorescently labeled cells in combination with data-driven agent-based modeling. The results indicate that just like WT cells, both mutants bias their movement toward aggregates and reduce motility inside aggregates. However, several aspects of mutant behavior remain uncorrected by WT, demonstrating that perfect recreation of WT behavior is unnecessary. In fact, synergies between errant behaviors can make aggregation robust. IMPORTANCE Self-organization into spatial patterns is evident in many multicellular phenomena. Even for the best-studied systems, our ability to dissect the mechanisms driving coordinated cell movement is limited. While genetic approaches can identify mutations perturbing multicellular patterns, the diverse nature of the signaling cues coupled to significant heterogeneity of individual cell behavior impedes our ability to mechanistically connect genes with phenotype. Small differences in the behaviors of mutant strains could be irrelevant or could sometimes lead to large differences in the emergent patterns. Here, we investigate rescue of multicellular aggregation in two mutant strains of Myxococcus xanthus mixed with wild-type cells. The results demonstrate how careful quantification of cell behavior coupled to data-driven modeling can identify specific motility features responsible for cell aggregation and thereby reveal important synergies and compensatory mechanisms. Notably, mutant cells do not need to precisely recreate wild-type behaviors to achieve complete aggregation. 
    more » « less