{"id":{"repo_id":"cambridge","oai_identifier":"oai:www.repository.cam.ac.uk:1810/375355"},"canonical_url":"https://search.dev.ndltd.org/etd/cambridge/oai:www.repository.cam.ac.uk:1810/375355","repository":{"repo_id":"cambridge","name":"Cambridge University","base_url":"https://api.repository.cam.ac.uk/server/oai/request"},"display":{"title":"Next-Generation Mendelian Randomization: Advanced and Reliable Methods for Complex Causal Inference","abstract":"Mendelian randomization is an epidemiological method that uses genetic variants as instrumental variables to study the causal effects of exposures on outcomes. Conventional MR is primarily implemented to test or estimate effects in relatively simple forms. However, to gain deeper insights into causal mechanisms, improve decision-making, and enhance the interpretation of results, more detailed effect forms are encouraged to be explored. This thesis extends conventional Mendelian randomization methods and proposes a series of novel, advanced approaches for studying more complex causal effect forms across various effect topics of potential interest in real applications. The thesis begins with two chapters that present the foundational topics of causal inference and Mendelian randomization. The novel work is then organized into three separate but interrelated chapters, each focusing on nonlinear effects, heterogeneous effects, and time-varying effects, respectively. For the nonlinear effect, we introduce the concept of stratification and a nonparametric stratification method, and propose advanced smoothing strategies to estimate potentially nonlinear or complex effect shapes. For the heterogeneous effect, we develop data-adaptive methods to investigate effect heterogeneity with high-dimensional covariates. We provide methods to test effect homogeneity, detect key effect drivers, and predict causal effects using individual covariate information. For the time-varying effect, we emphasize the importance of carefully considering time information in Mendelian randomization and explore continuous-time modelling. We present methods for estimating time-varying effects using the functional dimension reduction idea and combine them with identification-robust techniques. For each effect scenario, we apply our proposed methods to investigate the corresponding complex causal effect of a commonly-used exposure on a commonly-used outcome using data from the UK Biobank. The methods proposed in this thesis can be applied to estimate more complex effects or integrated into the toolbox of current Mendelian randomization to assess underlying assumptions, leading to more reliable conclusions. Software packages or code, along with guidance, are provided to implement all the proposed methods.","abstract_html":"Mendelian randomization is an epidemiological method that uses genetic variants as instrumental variables to study the causal effects of exposures on outcomes. Conventional MR is primarily implemented to test or estimate effects in relatively simple forms. However, to gain deeper insights into causal mechanisms, improve decision-making, and enhance the interpretation of results, more detailed effect forms are encouraged to be explored. This thesis extends conventional Mendelian randomization methods and proposes a series of novel, advanced approaches for studying more complex causal effect forms across various effect topics of potential interest in real applications. The thesis begins with two chapters that present the foundational topics of causal inference and Mendelian randomization. The novel work is then organized into three separate but interrelated chapters, each focusing on nonlinear effects, heterogeneous effects, and time-varying effects, respectively. For the nonlinear effect, we introduce the concept of stratification and a nonparametric stratification method, and propose advanced smoothing strategies to estimate potentially nonlinear or complex effect shapes. For the heterogeneous effect, we develop data-adaptive methods to investigate effect heterogeneity with high-dimensional covariates. We provide methods to test effect homogeneity, detect key effect drivers, and predict causal effects using individual covariate information. For the time-varying effect, we emphasize the importance of carefully considering time information in Mendelian randomization and explore continuous-time modelling. We present methods for estimating time-varying effects using the functional dimension reduction idea and combine them with identification-robust techniques. For each effect scenario, we apply our proposed methods to investigate the corresponding complex causal effect of a commonly-used exposure on a commonly-used outcome using data from the UK Biobank. The methods proposed in this thesis can be applied to estimate more complex effects or integrated into the toolbox of current Mendelian randomization to assess underlying assumptions, leading to more reliable conclusions. Software packages or code, along with guidance, are provided to implement all the proposed methods.","abstract_has_math":false,"creators":["Tian, Haodong"],"institution":"University of Cambridge","degree_name":"Doctor of Philosophy (PhD)","degree_level":"Doctoral","degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["Burgess, Stephen"],"committee_chairs":[],"committee_members":[],"year":2024,"date_issued":"2024-05-22","date_published":"2024-05-22","updated_at":"2026-07-22T22:23:54Z","subjects":["Causal inference","Effect heterogeneity","Effect shape","Instrumental variables","Mendelian randomization","Time-varying effect"],"languages":["eng"],"rights":[],"rights_urls":["https://www.repository.cam.ac.uk/bitstreams/1bcf3de2-423a-4c5f-899c-d18d8173c9d0/download","https://creativecommons.org/licenses/by-nc/4.0/"],"identifier_entries":[]},"links":{"outbound_url":"https://doi.org/10.17863/CAM.113069","outbound_label":"DOI","outbound_source":"dc:identifier.doi"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Burgess, Stephen"]},{"key":"dc:creator","label":"Author","values":["Tian, Haodong"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.issued","label":"Date","values":["2024-05-22"]},{"key":"dc:publisher.institution","label":"Dc Publisher Institution","values":["University of Cambridge"]},{"key":"dc:relation.isreferencedby.uri","label":"Dc Relation Isreferencedby URI","values":["https://www.repository.cam.ac.uk/handle/1810/375355"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"dc:type.qualificationlevel","label":"Dc Type Qualificationlevel","values":["Doctoral"]},{"key":"dc:type.qualificationname","label":"Dc Type Qualificationname","values":["Doctor of Philosophy (PhD)"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Causal inference","Effect heterogeneity","Effect shape","Instrumental variables","Mendelian randomization","Time-varying effect"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["https://www.repository.cam.ac.uk/bitstreams/1bcf3de2-423a-4c5f-899c-d18d8173c9d0/download","https://creativecommons.org/licenses/by-nc/4.0/"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.doi","label":"DOI","values":["https://doi.org/10.17863/CAM.113069"]},{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://www.repository.cam.ac.uk/bitstreams/37e4c46b-8683-4c94-825f-ad93b8bcbf75/download"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Mendelian randomization is an epidemiological method that uses genetic variants as instrumental variables to study the causal effects of exposures on outcomes. Conventional MR is primarily implemented to test or estimate effects in relatively simple forms. However, to gain deeper insights into causal mechanisms, improve decision-making, and enhance the interpretation of results, more detailed effect forms are encouraged to be explored. This thesis extends conventional Mendelian randomization methods and proposes a series of novel, advanced approaches for studying more complex causal effect forms across various effect topics of potential interest in real applications. The thesis begins with two chapters that present the foundational topics of causal inference and Mendelian randomization. The novel work is then organized into three separate but interrelated chapters, each focusing on nonlinear effects, heterogeneous effects, and time-varying effects, respectively. For the nonlinear effect, we introduce the concept of stratification and a nonparametric stratification method, and propose advanced smoothing strategies to estimate potentially nonlinear or complex effect shapes. For the heterogeneous effect, we develop data-adaptive methods to investigate effect heterogeneity with high-dimensional covariates. We provide methods to test effect homogeneity, detect key effect drivers, and predict causal effects using individual covariate information. For the time-varying effect, we emphasize the importance of carefully considering time information in Mendelian randomization and explore continuous-time modelling. We present methods for estimating time-varying effects using the functional dimension reduction idea and combine them with identification-robust techniques. For each effect scenario, we apply our proposed methods to investigate the corresponding complex causal effect of a commonly-used exposure on a commonly-used outcome using data from the UK Biobank. The methods proposed in this thesis can be applied to estimate more complex effects or integrated into the toolbox of current Mendelian randomization to assess underlying assumptions, leading to more reliable conclusions. Software packages or code, along with guidance, are provided to implement all the proposed methods."]},{"key":"dc:format.checksum.md5","label":"Dc Format Checksum Md5","values":["d791aa2909b6250308128bba44da6e7f","87eda9de84448d1f82354d60eee3eb5f"]},{"key":"dc:title","label":"Title","values":["Next-Generation Mendelian Randomization: Advanced and Reliable Methods for Complex Causal Inference"]}]}],"canonical_facts":{"dc:contributor.advisor":["Burgess, Stephen"],"dc:creator":["Tian, Haodong"],"dc:date.issued":["2024-05-22"],"dc:description.abstract":["Mendelian randomization is an epidemiological method that uses genetic variants as instrumental variables to study the causal effects of exposures on outcomes. Conventional MR is primarily implemented to test or estimate effects in relatively simple forms. However, to gain deeper insights into causal mechanisms, improve decision-making, and enhance the interpretation of results, more detailed effect forms are encouraged to be explored. This thesis extends conventional Mendelian randomization methods and proposes a series of novel, advanced approaches for studying more complex causal effect forms across various effect topics of potential interest in real applications. The thesis begins with two chapters that present the foundational topics of causal inference and Mendelian randomization. The novel work is then organized into three separate but interrelated chapters, each focusing on nonlinear effects, heterogeneous effects, and time-varying effects, respectively. For the nonlinear effect, we introduce the concept of stratification and a nonparametric stratification method, and propose advanced smoothing strategies to estimate potentially nonlinear or complex effect shapes. For the heterogeneous effect, we develop data-adaptive methods to investigate effect heterogeneity with high-dimensional covariates. We provide methods to test effect homogeneity, detect key effect drivers, and predict causal effects using individual covariate information. For the time-varying effect, we emphasize the importance of carefully considering time information in Mendelian randomization and explore continuous-time modelling. We present methods for estimating time-varying effects using the functional dimension reduction idea and combine them with identification-robust techniques. For each effect scenario, we apply our proposed methods to investigate the corresponding complex causal effect of a commonly-used exposure on a commonly-used outcome using data from the UK Biobank. The methods proposed in this thesis can be applied to estimate more complex effects or integrated into the toolbox of current Mendelian randomization to assess underlying assumptions, leading to more reliable conclusions. Software packages or code, along with guidance, are provided to implement all the proposed methods."],"dc:format.checksum.md5":["d791aa2909b6250308128bba44da6e7f","87eda9de84448d1f82354d60eee3eb5f"],"dc:identifier.doi":["https://doi.org/10.17863/CAM.113069"],"dc:identifier.uri":["https://www.repository.cam.ac.uk/bitstreams/37e4c46b-8683-4c94-825f-ad93b8bcbf75/download"],"dc:language":["eng"],"dc:publisher.institution":["University of Cambridge"],"dc:relation.isreferencedby.uri":["https://www.repository.cam.ac.uk/handle/1810/375355"],"dc:rights":["https://www.repository.cam.ac.uk/bitstreams/1bcf3de2-423a-4c5f-899c-d18d8173c9d0/download","https://creativecommons.org/licenses/by-nc/4.0/"],"dc:subject":["Causal inference","Effect heterogeneity","Effect shape","Instrumental variables","Mendelian randomization","Time-varying effect"],"dc:title":["Next-Generation Mendelian Randomization: Advanced and Reliable Methods for Complex Causal Inference"],"dc:type":["Thesis"],"dc:type.qualificationlevel":["Doctoral"],"dc:type.qualificationname":["Doctor of Philosophy (PhD)"]},"updated_at":"2026-07-22T22:23:54Z"}