{"id":{"repo_id":"rice","oai_identifier":"oai:repository.rice.edu:1911/118433"},"canonical_url":"https://search.dev.ndltd.org/etd/rice/oai:repository.rice.edu:1911/118433","repository":{"repo_id":"rice","name":"Rice University","base_url":"https://repository.rice.edu/server/oai/request"},"display":{"title":"Using Deep Learning to Extract Multicellular Aggregation Features of Myxococcus xanthus","abstract":"Self-organization is a fundamental biological process whereby local interactions among individual elements give rise to emergent global behaviors and complex patterns. While widely observed, the genetic determinants and biophysical mechanisms underlying self-organization—especially in prokaryotic multicellularity—remain incompletely understood. The bacterium Myxococcus xanthus, known for its intricate aggregation and fruiting body formation under nutrient deprivation, offers a powerful model for studying these dynamics. This thesis addresses both methodological and analytical challenges in studying M. xanthus development. To overcome the limitations of traditional imaging (e.g., difficulty in segmenting dense aggregates from phase-contrast alone), we developed a generative adversarial network (GAN) that synthesizes fluorescence microscopy images from phase-contrast images. This approach combines the ease of phase-contrast acquisition with the quantitative advantages of fluorescence imaging, improving aggregate segmentation and enabling accurate analysis of multicellular patterns such as rippling waves. We also introduce a deep learning framework for quantitative phenotypic analysis, integrating ResNet and StyleGAN2 into a Variational AutoEncoders (VAEs), and using Siamese architectures as the similarity metric. This pipeline transforms high-resolution microscopy data into low-dimensional phenotypic feature vectors. Human evaluations confirmed the model captures phenotypic diversity, enabling precise strain-level comparison of developmental dynamics. Altogether, this work demonstrates how deep learning can uncover genotype–phenotype relationships in bacterial self-organization and offers broadly applicable tools for the study of emergent behavior in biological systems.","abstract_html":"Self-organization is a fundamental biological process whereby local interactions among individual elements give rise to emergent global behaviors and complex patterns. While widely observed, the genetic determinants and biophysical mechanisms underlying self-organization—especially in prokaryotic multicellularity—remain incompletely understood. The bacterium Myxococcus xanthus, known for its intricate aggregation and fruiting body formation under nutrient deprivation, offers a powerful model for studying these dynamics. This thesis addresses both methodological and analytical challenges in studying M. xanthus development. To overcome the limitations of traditional imaging (e.g., difficulty in segmenting dense aggregates from phase-contrast alone), we developed a generative adversarial network (GAN) that synthesizes fluorescence microscopy images from phase-contrast images. This approach combines the ease of phase-contrast acquisition with the quantitative advantages of fluorescence imaging, improving aggregate segmentation and enabling accurate analysis of multicellular patterns such as rippling waves. We also introduce a deep learning framework for quantitative phenotypic analysis, integrating ResNet and StyleGAN2 into a Variational AutoEncoders (VAEs), and using Siamese architectures as the similarity metric. This pipeline transforms high-resolution microscopy data into low-dimensional phenotypic feature vectors. Human evaluations confirmed the model captures phenotypic diversity, enabling precise strain-level comparison of developmental dynamics. Altogether, this work demonstrates how deep learning can uncover genotype–phenotype relationships in bacterial self-organization and offers broadly applicable tools for the study of emergent behavior in biological systems.","abstract_has_math":false,"creators":["Zhang, Jiangguo"],"institution":"Rice University","degree_name":"Doctor of Philosophy","degree_level":"Doctoral","degree_discipline":"Engineering","degree_department":null,"school":null,"contributors":[],"advisors":["Igoshin, Oleg A."],"committee_chairs":[],"committee_members":[],"year":2025,"date_issued":"2025-04-25","date_published":"2025-04-25","updated_at":"2026-07-24T04:10:34Z","subjects":["Myxococcus xanthus","phase contrast microscopy","fluorescence microscopy","aggregation","rippling","deep learning","generative adversarial network","ResNet","StyleGAN","Variational Auto-encoder","Siamese network, change point detection"],"languages":["eng"],"rights":["Copyright is held by the author, unless otherwise indicated. Permission to reuse, publish, or reproduce the work beyond the bounds of fair use or other exemptions to copyright law must be obtained from the copyright holder."],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/1911/118433","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Igoshin, Oleg A."]},{"key":"dc:creator","label":"Author","values":["Zhang, Jiangguo"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2025-05-29T19:56:38Z"]},{"key":"dc:date.issued","label":"Date","values":["2025-04-25"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Engineering"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Doctoral"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Doctor of Philosophy"]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["Rice University"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Myxococcus xanthus","phase contrast microscopy","fluorescence microscopy","aggregation","rippling","deep learning","generative adversarial network","ResNet","StyleGAN","Variational Auto-encoder","Siamese network, change point detection"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright is held by the author, unless otherwise indicated. Permission to reuse, publish, or reproduce the work beyond the bounds of fair use or other exemptions to copyright law must be obtained from the copyright holder."]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/1911/118433"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Self-organization is a fundamental biological process whereby local interactions among individual elements give rise to emergent global behaviors and complex patterns. While widely observed, the genetic determinants and biophysical mechanisms underlying self-organization—especially in prokaryotic multicellularity—remain incompletely understood. The bacterium Myxococcus xanthus, known for its intricate aggregation and fruiting body formation under nutrient deprivation, offers a powerful model for studying these dynamics. This thesis addresses both methodological and analytical challenges in studying M. xanthus development. To overcome the limitations of traditional imaging (e.g., difficulty in segmenting dense aggregates from phase-contrast alone), we developed a generative adversarial network (GAN) that synthesizes fluorescence microscopy images from phase-contrast images. This approach combines the ease of phase-contrast acquisition with the quantitative advantages of fluorescence imaging, improving aggregate segmentation and enabling accurate analysis of multicellular patterns such as rippling waves. We also introduce a deep learning framework for quantitative phenotypic analysis, integrating ResNet and StyleGAN2 into a Variational AutoEncoders (VAEs), and using Siamese architectures as the similarity metric. This pipeline transforms high-resolution microscopy data into low-dimensional phenotypic feature vectors. Human evaluations confirmed the model captures phenotypic diversity, enabling precise strain-level comparison of developmental dynamics. Altogether, this work demonstrates how deep learning can uncover genotype–phenotype relationships in bacterial self-organization and offers broadly applicable tools for the study of emergent behavior in biological systems."]},{"key":"dc:format.mimetype","label":"Dc Format Mimetype","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Using Deep Learning to Extract Multicellular Aggregation Features of Myxococcus xanthus"]}]}],"canonical_facts":{"dc:contributor.advisor":["Igoshin, Oleg A."],"dc:creator":["Zhang, Jiangguo"],"dc:date.accessioned":["2025-05-29T19:56:38Z"],"dc:date.issued":["2025-04-25"],"dc:description.abstract":["Self-organization is a fundamental biological process whereby local interactions among individual elements give rise to emergent global behaviors and complex patterns. While widely observed, the genetic determinants and biophysical mechanisms underlying self-organization—especially in prokaryotic multicellularity—remain incompletely understood. The bacterium Myxococcus xanthus, known for its intricate aggregation and fruiting body formation under nutrient deprivation, offers a powerful model for studying these dynamics. This thesis addresses both methodological and analytical challenges in studying M. xanthus development. To overcome the limitations of traditional imaging (e.g., difficulty in segmenting dense aggregates from phase-contrast alone), we developed a generative adversarial network (GAN) that synthesizes fluorescence microscopy images from phase-contrast images. This approach combines the ease of phase-contrast acquisition with the quantitative advantages of fluorescence imaging, improving aggregate segmentation and enabling accurate analysis of multicellular patterns such as rippling waves. We also introduce a deep learning framework for quantitative phenotypic analysis, integrating ResNet and StyleGAN2 into a Variational AutoEncoders (VAEs), and using Siamese architectures as the similarity metric. This pipeline transforms high-resolution microscopy data into low-dimensional phenotypic feature vectors. Human evaluations confirmed the model captures phenotypic diversity, enabling precise strain-level comparison of developmental dynamics. Altogether, this work demonstrates how deep learning can uncover genotype–phenotype relationships in bacterial self-organization and offers broadly applicable tools for the study of emergent behavior in biological systems."],"dc:format.mimetype":["application/pdf"],"dc:identifier.uri":["https://hdl.handle.net/1911/118433"],"dc:language.iso":["eng"],"dc:rights":["Copyright is held by the author, unless otherwise indicated. Permission to reuse, publish, or reproduce the work beyond the bounds of fair use or other exemptions to copyright law must be obtained from the copyright holder."],"dc:subject":["Myxococcus xanthus","phase contrast microscopy","fluorescence microscopy","aggregation","rippling","deep learning","generative adversarial network","ResNet","StyleGAN","Variational Auto-encoder","Siamese network, change point detection"],"dc:title":["Using Deep Learning to Extract Multicellular Aggregation Features of Myxococcus xanthus"],"dc:type":["Thesis"],"thesis:degree_discipline":["Engineering"],"thesis:degree_level":["Doctoral"],"thesis:degree_name":["Doctor of Philosophy"],"thesis:institution_name":["Rice University"]},"updated_at":"2026-07-24T04:10:34Z"}