{"id":{"repo_id":"msu","oai_identifier":"oai:d.lib.msu.edu:etd_53383"},"canonical_url":"https://search.ndltd.org/etd/msu/oai:d.lib.msu.edu:etd_53383","repository":{"repo_id":"msu","name":"Michigan State University","base_url":"https://d.lib.msu.edu/oai"},"display":{"title":"STABILIZING SPATIAL TRANSCRIPTOMIC CLUSTERING USING RADIOMICS-DERIVED HISTOLOGICAL FEATURES","abstract":"Thesis (Ph.D.)--Michigan State University. Computational Mathematics, Science and Engineering - Doctor of Philosophy, 2026","abstract_html":"Thesis (Ph.D.)--Michigan State University. Computational Mathematics, Science and Engineering - Doctor of Philosophy, 2026","abstract_has_math":false,"creators":["Loveless, Ian M."],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Alessio, Adam","Chen, Bin","Cao, Guanqun","Wang, Jianrong"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026","date_published":"2026","updated_at":"2026-11-18T18:59:47Z","subjects":["Bioinformatics"],"languages":["eng"],"rights":["Attribution-NoDerivatives 4.0 International"],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:53383","local:Loveless_grad.msu_0128D_22440"],"render_values":[{"text":"etd:53383","href":null,"code":true},{"text":"local:Loveless_grad.msu_0128D_22440","href":null,"code":true}]}],"person_identifier_entries":[]},"links":{"outbound_url":"https://doi.org/doi:10.25335/8qzj-bd14","outbound_label":"DOI","outbound_source":"dc:identifier","record_url":"https://doi.org/doi:10.25335/8qzj-bd14","record_label":"Repository record","record_kind":"doi","record_source":"dc:identifier","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://d.lib.msu.edu/oai?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3Ad.lib.msu.edu%3Aetd_53383","prefix":"oai_dc"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Alessio, Adam","Chen, Bin","Cao, Guanqun","Wang, Jianrong"]},{"key":"dc:creator","label":"Author","values":["Loveless, Ian M."]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2026"]},{"key":"dc:relation","label":"Dc Relation","values":["Electronic Theses & Dissertations"]},{"key":"dc:type","label":"Dc Type","values":["text","Electronic dissertations"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Bioinformatics"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["Attribution-NoDerivatives 4.0 International"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:53383","local:Loveless_grad.msu_0128D_22440","https://doi.org/doi:10.25335/8qzj-bd14"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Thesis (Ph.D.)--Michigan State University. Computational Mathematics, Science and Engineering - Doctor of Philosophy, 2026","Spatial transcriptomics enables gene expression to be measured within intact tissue whilepreserving the spatial organization necessary to characterize tissue architecture and cellular microenvironments. However, the reliability of spatial clustering is fundamentally limited by the quality of the underlying transcriptomic measurements. Reduced capture efficiency, ambient RNA contamination, and transcript dropout progressively distort transcriptomic representations, leading to unstable neighborhood graphs and loss of biologically meaningful spatial domains. These limitations are particularly important for formalin-fixed paraffin- embedded tissues, archival clinical specimens, and other challenging experimental settings in which transcriptomic degradation is common. This dissertation investigates the robustness of spatial transcriptomic clustering under controlled transcriptomic degradation and develops a representation-centered framework for improving clustering stability through multimodal integration. We first introduce a math- ematical framework that models transcriptomic degradation through reduced capture effi- ciency, ambient RNA contamination, and transcript dropout, enabling systematic evaluation of clustering robustness across multiple degradation regimes. Using complementary mea- sures of representation geometry, neighborhood preservation, graph connectivity, biological fidelity, and clustering stability, we demonstrate that transcriptomic degradation produces a consistent failure cascade in which distortion of the learned representation rewires local neighborhood relationships, fragments graph topology, and ultimately destabilizes graph- based clustering. To mitigate these effects, we introduce HIPSTR, a multimodal framework that inte- grates quantitative radiomic features extracted from co-registered histological images with transcriptomic measurements. Because tissue morphology remains largely invariant to tran- scriptomic degradation, morphology-derived features provide complementary structural in- formation that stabilizes the learned representation when transcriptomic similarity becomes unreliable. Multiple multimodal fusion strategies were evaluated across independent datasets and degradation conditions, demonstrating consistent improvements in clustering robustness relative to RNA-only approaches. Collectively, this work establishes that robust spatial transcriptomic analysis depends fundamentally on stable representation learning rather than clustering algorithms alone. By integrating complementary molecular and morphological information, the proposed frame- work preserves biologically meaningful spatial organization under realistic experimental per- turbations and extends the applicability of spatial transcriptomics to noisy, degraded, and archival tissue samples.","Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).","Includes bibliographical references."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf","1 online resource (177 p.) : ill."]},{"key":"dc:title","label":"Title","values":["STABILIZING SPATIAL TRANSCRIPTOMIC CLUSTERING USING RADIOMICS-DERIVED HISTOLOGICAL FEATURES"]}]}],"canonical_facts":{"dc:contributor":["Alessio, Adam","Chen, Bin","Cao, Guanqun","Wang, Jianrong"],"dc:creator":["Loveless, Ian M."],"dc:date":["2026"],"dc:description":["Thesis (Ph.D.)--Michigan State University. Computational Mathematics, Science and Engineering - Doctor of Philosophy, 2026","Spatial transcriptomics enables gene expression to be measured within intact tissue whilepreserving the spatial organization necessary to characterize tissue architecture and cellular microenvironments. However, the reliability of spatial clustering is fundamentally limited by the quality of the underlying transcriptomic measurements. Reduced capture efficiency, ambient RNA contamination, and transcript dropout progressively distort transcriptomic representations, leading to unstable neighborhood graphs and loss of biologically meaningful spatial domains. These limitations are particularly important for formalin-fixed paraffin- embedded tissues, archival clinical specimens, and other challenging experimental settings in which transcriptomic degradation is common. This dissertation investigates the robustness of spatial transcriptomic clustering under controlled transcriptomic degradation and develops a representation-centered framework for improving clustering stability through multimodal integration. We first introduce a math- ematical framework that models transcriptomic degradation through reduced capture effi- ciency, ambient RNA contamination, and transcript dropout, enabling systematic evaluation of clustering robustness across multiple degradation regimes. Using complementary mea- sures of representation geometry, neighborhood preservation, graph connectivity, biological fidelity, and clustering stability, we demonstrate that transcriptomic degradation produces a consistent failure cascade in which distortion of the learned representation rewires local neighborhood relationships, fragments graph topology, and ultimately destabilizes graph- based clustering. To mitigate these effects, we introduce HIPSTR, a multimodal framework that inte- grates quantitative radiomic features extracted from co-registered histological images with transcriptomic measurements. Because tissue morphology remains largely invariant to tran- scriptomic degradation, morphology-derived features provide complementary structural in- formation that stabilizes the learned representation when transcriptomic similarity becomes unreliable. Multiple multimodal fusion strategies were evaluated across independent datasets and degradation conditions, demonstrating consistent improvements in clustering robustness relative to RNA-only approaches. Collectively, this work establishes that robust spatial transcriptomic analysis depends fundamentally on stable representation learning rather than clustering algorithms alone. By integrating complementary molecular and morphological information, the proposed frame- work preserves biologically meaningful spatial organization under realistic experimental per- turbations and extends the applicability of spatial transcriptomics to noisy, degraded, and archival tissue samples.","Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).","Includes bibliographical references."],"dc:format":["application/pdf","1 online resource (177 p.) : ill."],"dc:identifier":["etd:53383","local:Loveless_grad.msu_0128D_22440","https://doi.org/doi:10.25335/8qzj-bd14"],"dc:language":["eng"],"dc:relation":["Electronic Theses & Dissertations"],"dc:rights":["Attribution-NoDerivatives 4.0 International"],"dc:subject":["Bioinformatics"],"dc:title":["STABILIZING SPATIAL TRANSCRIPTOMIC CLUSTERING USING RADIOMICS-DERIVED HISTOLOGICAL FEATURES"],"dc:type":["text","Electronic dissertations"]},"updated_at":"2026-11-18T18:59:47Z"}