There are many active research projects accessing and applying shared ADNI data. Use the search above to find specific research focuses on the active ADNI investigations. This information is requested annually as a requirement for data access.
| Principal Investigator | |
| Principal Investigator's Name: | John Williams |
| Institution: | University of Birmingham |
| Department: | Institute of Cancer and Genomic Sciences |
| Country: | |
| Proposed Analysis: | Aims: Our aim in this study is to: generate a semantic in-depth phenotypic characterization of ANDI subjects and use this to stratify patients into sub-cohorts. We will apply an AI-based multimodal integration analysis framework which will utilize the existing genomic and imaging data to extract epigenetic and combinatorial networks of features which can segregate patients into sub-cohorts via predictive machine learning techniques. Lastly, we aim to use explanatory omics/imaging derived feature networks to interrogate subjects’ underlying biology; and gain an in-depth understanding of the pathophysiology of heterogeneous patients with diagnosed Alzheimer’s disease. Background: For more than a decade, we have been developing approaches to the formulation of formal definitions of phenotypes and associated data sets, which have proved key to our ability to develop methods for the integration and analysis of phenotype data. Our group has developed approaches which allow us to distinguish which of the many thousands of DNA sequence variants carried by an individual patient are responsible for the disease-associated phenotypes that patient presents with the PhenomeNet ontology and recent PhenomeNET variant predictor (1,2). These approaches have been built around behavior-related manifestations and disorders. We have developed the Neuro Behavior Ontology, the international standard for the formal description of behavior and its associated manifestation that caters to the integration of genetic and behavior-related data (3). We have applied our approach to a variety of applications, including behavioral annotation of gene expression (4). Existing Alzheimer’s disease ontologies exist but focus entirely on genomics and biochemical pathways (5) or are not integrated into existing biomedical ontologies (6), revealing a space for an integrative ontological approach to Alzheimer’s patient description and stratification. Methods: Our methods rely on feature mining and engineering from three domains of data available from the ADNI cohorts: genomic/proteomic, brain imaging, and clinical phenotyping. In each domain there exist heterogenous patterns amongst both controls and diagnosed subjects, which we will leverage to reveal diagnostically relevant patient profiles. We intend to identify causative gene and structural networks derived from integrated analysis of data available in the ADNI cohorts, focusing on structural, diffusion tensor, and functional MRI imaging data and whole genome and SNP array genomics data which influence aspects of Alzheimer’s disease as measured by clinical assessment. To build patient profiles, we will incorporate ADNI clinical measures into biomedical ontologies which model logical relationships between biomedical traits presented by people with Alzheimer’s disease and cognitive impairment. Existing resources used to model traits include the Human Phenotype (7), Neuro Behavior (3), and Alzheimer’s Disease (6) ontologies. Once profiles are created, unsupervised learning methods (hierarchical, spectral, neural network-based clustering) will be created from ontologies annotated with subject data. Semantic similarity measures (8) will be applied to create numerical patient profiles representing clinic-derived subtypes of subjects with Alzheimer’s disease. To find biomarkers that represent these patient subtypes, association analyses will be performed with associate genomic and imaging networks, using patient profiles as targets in supervised learning. For each genomics and imaging dataset, we will independently apply feature selection algorithms which have been successful in literature to extract features relevant for each data type (9). In the structural imaging domain, feature engineering will include additive and multiplicative relationships between brain region sizes (10). In the genomics domain, we will be guided by the omnigenic hypothesis (11) and test for associations between epistatic networks of 2nd and 3rd order gene/gene interactions. To select features from such high dimensional space, we will include both linear (12,13) and non-linear (14,15) methods. Features selected for each modality will be combined and input into existing machine learning pipelines our group has implemented for precision medicine applications (16). Briefly, these methods combine extensive iterative resampling techniques with stochastic methods (random forest, neural networks, extreme gradient boosting machines) to create multiple models which use the same features in differing ways and extract features which dominate multiple model types. These are then combined in an ensemble-learning approach to enhance predictive modelling. This approach will be used to model both Alzheimer’s subtypes and the stages of Alzheimer’s and cognitive decline previously analysed in the ADNI datasets. In addition to this hypothesis-free, data-driven features selection approach, genomic signals from circadian and sleep-associated loci will be independently investigated and associated with other features in a planned sub-analysis. Anticipated Results: Our anticipated results will follow four strands of the analysis described above. Using patient-level information presented in clinical and psychiatric assessments available in the ADNI resource, we will project patient characteristics into the phenotype space by integrating features from several well-developed biomedical ontologies, potentially creating a new depiction of Alzheimer’s symptoms. Importantly, profiles of subjects with similar clinical features will be created, leading to patient subgroups for further structural and molecular characterization. Genomic analyses of these cohorts will leverage tests for epistatic interactions between genes, which will create gene-level networks which help characterize ADNI subpopulations. Features derived from structural MRI will likewise be created, and features which discriminate between phenotype-driven subtypes will be investigated. The combination of genomic and imaging features for each subtype of ADNI subjects identified will be tested together, and their ability to classify subjects into correct cohorts will be thoroughly evaluated. Lastly, correlations between features themselves will be investigated, focusing on circadian-related genomic loci and their relationship to structural morphology in ADNI subjects. This will, ideally, evaluate the relationship between loci associated with abnormal circadian biology, cognitive decline, and the structural brain biology of ADNI subpopulations. Conclusions: Our combination of patient-subgroup gene association with brain region-specific structural study will provide a novel set of potential biomarkers (features) to subset a population with a diverse phenotypic profile. While basic aims in this proposed work include method patient characterization and multiple machine learning approaches applied to many datasets, the ultimate aim is to improve knowledge of which genetic lesions and brain morphological characteristics contributing to Alzheimer’s patient phenotypes. References: 1. Hoehndorf R, Schofield PN, Gkoutos GV. PhenomeNET: a whole-phenome approach to disease gene discovery. Nucleic Acids Res. 2011 Oct 1;39(18):e119–e119. 2. Boudellioua I, Mahamad Razali RB, Kulmanov M, Hashish Y, Bajic VB, Goncalves-Serra E, et al. Semantic prioritization of novel causative genomic variants. PLoS Comput Biol. 2017 Apr;13(4):e1005500. 3. Gkoutos GV, Schofield PN, Hoehndorf R. The neurobehavior ontology: an ontology for annotation and integration of behavior and behavioral phenotypes. Int Rev Neurobiol. 2012;103:69–87. 4. Hoehndorf R, Hancock JM, Hardy NW, Mallon A-M, Schofield PN, Gkoutos GV. Analyzing gene expression data in mice with the Neuro Behavior Ontology. Mammalian Genome. 2014 Feb;25(1–2):32–40. 5. Henry V, Moszer I, Dameron O, Potier M-C, Hofmann-Apitius M, Colliot O. Converting Alzheimer s disease map into a heavyweight ontology: a formal network to integrate data. arXiv:180710509 [q-bio] [Internet]. 2018 Jul 27 [cited 2019 Dec 13]; Available from: http://arxiv.org/abs/1807.10509 6. Malhotra A, Younesi E, Gündel M, Müller B, Heneka MT, Hofmann-Apitius M. ADO: a disease ontology representing the domain knowledge specific to Alzheimer’s disease. Alzheimers Dement. 2014 Mar;10(2):238–46. 7. Köhler S, Carmody L, Vasilevsky N, Jacobsen JOB, Danis D, Gourdine J-P, et al. Expansion of the Human Phenotype Ontology (HPO) knowledge base and resources. Nucleic Acids Res. 2019 Jan 8;47(D1):D1018–27. 8. Pesquita C. Semantic Similarity in the Gene Ontology. In: Dessimoz C, Škunca N, editors. The Gene Ontology Handbook [Internet]. New York, NY: Springer; 2017 [cited 2019 Dec 20]. p. 161–73. (Methods in Molecular Biology). Available from: https://doi.org/10.1007/978-1-4939-3743-1_12 9. Rathore S, Habes M, Iftikhar MA, Shacklett A, Davatzikos C. A review on neuroimaging-based classification studies and associated feature extraction methods for Alzheimer’s disease and its prodromal stages. NeuroImage. 2017 Jul 15;155:530–48. 10. Chincarini A, Bosco P, Calvini P, Gemme G, Esposito M, Olivieri C, et al. Local MRI analysis approach in the diagnosis of early and prodromal Alzheimer’s disease. NeuroImage. 2011 Sep 15;58(2):469–80. 11. Boyle EA, Li YI, Pritchard JK. An Expanded View of Complex Traits: From Polygenic to Omnigenic. Cell. 2017 Jun 15;169(7):1177–86. 12. Wu X, Dong H, Luo L, Zhu Y, Peng G, Reveille JD, et al. A Novel Statistic for Genome-Wide Interaction Analysis. PLOS Genetics. 2010 Sep 23;6(9):e1001131. 13. Ueki M, Cordell HJ. Improved Statistics for Genome-Wide Interaction Analysis. PLOS Genetics. 2012 Apr 5;8(4):e1002625. 14. Basu S, Kumbier K, Brown JB, Yu B. Iterative random forests to discover predictive and stable high-order interactions. PNAS. 2018 Feb 20;115(8):1943–8. 15. Kumbier K, Basu S, Brown JB, Celniker S, Yu B. Refining interaction search through signed iterative Random Forests. arXiv:181007287 [cs, stat] [Internet]. 2018 Oct 16 [cited 2019 May 10]; Available from: http://arxiv.org/abs/1810.07287 16. Bravo-Merodio L, Williams JA, Gkoutos GV, Acharjee A. -Omics biomarker identification pipeline for translational medicine. J Transl Med. 2019 May 14;17(1):155. |
| Additional Investigators | |
| Investigator's Name: | Georgios Gkoutos |
| Proposed Analysis: | Aims: Our aim in this study is to: generate a semantic in-depth phenotypic characterization of ANDI subjects and use this to stratify patients into sub-cohorts. We will apply an AI-based multimodal integration analysis framework which will utilize the existing genomic and imaging data to extract epigenetic and combinatorial networks of features which can segregate patients into sub-cohorts via predictive machine learning techniques. Lastly, we aim to use explanatory omics/imaging derived feature networks to interrogate subjects’ underlying biology; and gain an in-depth understanding of the pathophysiology of heterogeneous patients with diagnosed Alzheimer’s disease. Background: For more than a decade, we have been developing approaches to the formulation of formal definitions of phenotypes and associated data sets, which have proved key to our ability to develop methods for the integration and analysis of phenotype data. Our group has developed approaches which allow us to distinguish which of the many thousands of DNA sequence variants carried by an individual patient are responsible for the disease-associated phenotypes that patient presents with the PhenomeNet ontology and recent PhenomeNET variant predictor (1,2). These approaches have been built around behavior-related manifestations and disorders. We have developed the Neuro Behavior Ontology, the international standard for the formal description of behavior and its associated manifestation that caters to the integration of genetic and behavior-related data (3). We have applied our approach to a variety of applications, including behavioral annotation of gene expression (4). Existing Alzheimer’s disease ontologies exist but focus entirely on genomics and biochemical pathways (5) or are not integrated into existing biomedical ontologies (6), revealing a space for an integrative ontological approach to Alzheimer’s patient description and stratification. Methods: Our methods rely on feature mining and engineering from three domains of data available from the ADNI cohorts: genomic/proteomic, brain imaging, and clinical phenotyping. In each domain there exist heterogenous patterns amongst both controls and diagnosed subjects, which we will leverage to reveal diagnostically relevant patient profiles. We intend to identify causative gene and structural networks derived from integrated analysis of data available in the ADNI cohorts, focusing on structural, diffusion tensor, and functional MRI imaging data and whole genome and SNP array genomics data which influence aspects of Alzheimer’s disease as measured by clinical assessment. To build patient profiles, we will incorporate ADNI clinical measures into biomedical ontologies which model logical relationships between biomedical traits presented by people with Alzheimer’s disease and cognitive impairment. Existing resources used to model traits include the Human Phenotype (7), Neuro Behavior (3), and Alzheimer’s Disease (6) ontologies. Once profiles are created, unsupervised learning methods (hierarchical, spectral, neural network-based clustering) will be created from ontologies annotated with subject data. Semantic similarity measures (8) will be applied to create numerical patient profiles representing clinic-derived subtypes of subjects with Alzheimer’s disease. To find biomarkers that represent these patient subtypes, association analyses will be performed with associate genomic and imaging networks, using patient profiles as targets in supervised learning. For each genomics and imaging dataset, we will independently apply feature selection algorithms which have been successful in literature to extract features relevant for each data type (9). In the structural imaging domain, feature engineering will include additive and multiplicative relationships between brain region sizes (10). In the genomics domain, we will be guided by the omnigenic hypothesis (11) and test for associations between epistatic networks of 2nd and 3rd order gene/gene interactions. To select features from such high dimensional space, we will include both linear (12,13) and non-linear (14,15) methods. Features selected for each modality will be combined and input into existing machine learning pipelines our group has implemented for precision medicine applications (16). Briefly, these methods combine extensive iterative resampling techniques with stochastic methods (random forest, neural networks, extreme gradient boosting machines) to create multiple models which use the same features in differing ways and extract features which dominate multiple model types. These are then combined in an ensemble-learning approach to enhance predictive modelling. This approach will be used to model both Alzheimer’s subtypes and the stages of Alzheimer’s and cognitive decline previously analysed in the ADNI datasets. In addition to this hypothesis-free, data-driven features selection approach, genomic signals from circadian and sleep-associated loci will be independently investigated and associated with other features in a planned sub-analysis. Anticipated Results: Our anticipated results will follow four strands of the analysis described above. Using patient-level information presented in clinical and psychiatric assessments available in the ADNI resource, we will project patient characteristics into the phenotype space by integrating features from several well-developed biomedical ontologies, potentially creating a new depiction of Alzheimer’s symptoms. Importantly, profiles of subjects with similar clinical features will be created, leading to patient subgroups for further structural and molecular characterization. Genomic analyses of these cohorts will leverage tests for epistatic interactions between genes, which will create gene-level networks which help characterize ADNI subpopulations. Features derived from structural MRI will likewise be created, and features which discriminate between phenotype-driven subtypes will be investigated. The combination of genomic and imaging features for each subtype of ADNI subjects identified will be tested together, and their ability to classify subjects into correct cohorts will be thoroughly evaluated. Lastly, correlations between features themselves will be investigated, focusing on circadian-related genomic loci and their relationship to structural morphology in ADNI subjects. This will, ideally, evaluate the relationship between loci associated with abnormal circadian biology, cognitive decline, and the structural brain biology of ADNI subpopulations. Conclusions: Our combination of patient-subgroup gene association with brain region-specific structural study will provide a novel set of potential biomarkers (features) to subset a population with a diverse phenotypic profile. While basic aims in this proposed work include method patient characterization and multiple machine learning approaches applied to many datasets, the ultimate aim is to improve knowledge of which genetic lesions and brain morphological characteristics contributing to Alzheimer’s patient phenotypes. References: 1. Hoehndorf R, Schofield PN, Gkoutos GV. PhenomeNET: a whole-phenome approach to disease gene discovery. Nucleic Acids Res. 2011 Oct 1;39(18):e119–e119. 2. Boudellioua I, Mahamad Razali RB, Kulmanov M, Hashish Y, Bajic VB, Goncalves-Serra E, et al. Semantic prioritization of novel causative genomic variants. PLoS Comput Biol. 2017 Apr;13(4):e1005500. 3. Gkoutos GV, Schofield PN, Hoehndorf R. The neurobehavior ontology: an ontology for annotation and integration of behavior and behavioral phenotypes. Int Rev Neurobiol. 2012;103:69–87. 4. Hoehndorf R, Hancock JM, Hardy NW, Mallon A-M, Schofield PN, Gkoutos GV. Analyzing gene expression data in mice with the Neuro Behavior Ontology. Mammalian Genome. 2014 Feb;25(1–2):32–40. 5. Henry V, Moszer I, Dameron O, Potier M-C, Hofmann-Apitius M, Colliot O. Converting Alzheimer s disease map into a heavyweight ontology: a formal network to integrate data. arXiv:180710509 [q-bio] [Internet]. 2018 Jul 27 [cited 2019 Dec 13]; Available from: http://arxiv.org/abs/1807.10509 6. Malhotra A, Younesi E, Gündel M, Müller B, Heneka MT, Hofmann-Apitius M. ADO: a disease ontology representing the domain knowledge specific to Alzheimer’s disease. Alzheimers Dement. 2014 Mar;10(2):238–46. 7. Köhler S, Carmody L, Vasilevsky N, Jacobsen JOB, Danis D, Gourdine J-P, et al. Expansion of the Human Phenotype Ontology (HPO) knowledge base and resources. Nucleic Acids Res. 2019 Jan 8;47(D1):D1018–27. 8. Pesquita C. Semantic Similarity in the Gene Ontology. In: Dessimoz C, Škunca N, editors. The Gene Ontology Handbook [Internet]. New York, NY: Springer; 2017 [cited 2019 Dec 20]. p. 161–73. (Methods in Molecular Biology). Available from: https://doi.org/10.1007/978-1-4939-3743-1_12 9. Rathore S, Habes M, Iftikhar MA, Shacklett A, Davatzikos C. A review on neuroimaging-based classification studies and associated feature extraction methods for Alzheimer’s disease and its prodromal stages. NeuroImage. 2017 Jul 15;155:530–48. 10. Chincarini A, Bosco P, Calvini P, Gemme G, Esposito M, Olivieri C, et al. Local MRI analysis approach in the diagnosis of early and prodromal Alzheimer’s disease. NeuroImage. 2011 Sep 15;58(2):469–80. 11. Boyle EA, Li YI, Pritchard JK. An Expanded View of Complex Traits: From Polygenic to Omnigenic. Cell. 2017 Jun 15;169(7):1177–86. 12. Wu X, Dong H, Luo L, Zhu Y, Peng G, Reveille JD, et al. A Novel Statistic for Genome-Wide Interaction Analysis. PLOS Genetics. 2010 Sep 23;6(9):e1001131. 13. Ueki M, Cordell HJ. Improved Statistics for Genome-Wide Interaction Analysis. PLOS Genetics. 2012 Apr 5;8(4):e1002625. 14. Basu S, Kumbier K, Brown JB, Yu B. Iterative random forests to discover predictive and stable high-order interactions. PNAS. 2018 Feb 20;115(8):1943–8. 15. Kumbier K, Basu S, Brown JB, Celniker S, Yu B. Refining interaction search through signed iterative Random Forests. arXiv:181007287 [cs, stat] [Internet]. 2018 Oct 16 [cited 2019 May 10]; Available from: http://arxiv.org/abs/1810.07287 16. Bravo-Merodio L, Williams JA, Gkoutos GV, Acharjee A. -Omics biomarker identification pipeline for translational medicine. J Transl Med. 2019 May 14;17(1):155. |
| Investigator's Name: | Dominic Russ |
| Proposed Analysis: | Aims: Our aim in this study is to: generate a semantic in-depth phenotypic characterization of ANDI subjects and use this to stratify patients into sub-cohorts. We will apply an AI-based multimodal integration analysis framework which will utilize the existing genomic and imaging data to extract epigenetic and combinatorial networks of features which can segregate patients into sub-cohorts via predictive machine learning techniques. Lastly, we aim to use explanatory omics/imaging derived feature networks to interrogate subjects’ underlying biology; and gain an in-depth understanding of the pathophysiology of heterogeneous patients with diagnosed Alzheimer’s disease. Background: For more than a decade, we have been developing approaches to the formulation of formal definitions of phenotypes and associated data sets, which have proved key to our ability to develop methods for the integration and analysis of phenotype data. Our group has developed approaches which allow us to distinguish which of the many thousands of DNA sequence variants carried by an individual patient are responsible for the disease-associated phenotypes that patient presents with the PhenomeNet ontology and recent PhenomeNET variant predictor (1,2). These approaches have been built around behavior-related manifestations and disorders. We have developed the Neuro Behavior Ontology, the international standard for the formal description of behavior and its associated manifestation that caters to the integration of genetic and behavior-related data (3). We have applied our approach to a variety of applications, including behavioral annotation of gene expression (4). Existing Alzheimer’s disease ontologies exist but focus entirely on genomics and biochemical pathways (5) or are not integrated into existing biomedical ontologies (6), revealing a space for an integrative ontological approach to Alzheimer’s patient description and stratification. Methods: Our methods rely on feature mining and engineering from three domains of data available from the ADNI cohorts: genomic/proteomic, brain imaging, and clinical phenotyping. In each domain there exist heterogenous patterns amongst both controls and diagnosed subjects, which we will leverage to reveal diagnostically relevant patient profiles. We intend to identify causative gene and structural networks derived from integrated analysis of data available in the ADNI cohorts, focusing on structural, diffusion tensor, and functional MRI imaging data and whole genome and SNP array genomics data which influence aspects of Alzheimer’s disease as measured by clinical assessment. To build patient profiles, we will incorporate ADNI clinical measures into biomedical ontologies which model logical relationships between biomedical traits presented by people with Alzheimer’s disease and cognitive impairment. Existing resources used to model traits include the Human Phenotype (7), Neuro Behavior (3), and Alzheimer’s Disease (6) ontologies. Once profiles are created, unsupervised learning methods (hierarchical, spectral, neural network-based clustering) will be created from ontologies annotated with subject data. Semantic similarity measures (8) will be applied to create numerical patient profiles representing clinic-derived subtypes of subjects with Alzheimer’s disease. To find biomarkers that represent these patient subtypes, association analyses will be performed with associate genomic and imaging networks, using patient profiles as targets in supervised learning. For each genomics and imaging dataset, we will independently apply feature selection algorithms which have been successful in literature to extract features relevant for each data type (9). In the structural imaging domain, feature engineering will include additive and multiplicative relationships between brain region sizes (10). In the genomics domain, we will be guided by the omnigenic hypothesis (11) and test for associations between epistatic networks of 2nd and 3rd order gene/gene interactions. To select features from such high dimensional space, we will include both linear (12,13) and non-linear (14,15) methods. Features selected for each modality will be combined and input into existing machine learning pipelines our group has implemented for precision medicine applications (16). Briefly, these methods combine extensive iterative resampling techniques with stochastic methods (random forest, neural networks, extreme gradient boosting machines) to create multiple models which use the same features in differing ways and extract features which dominate multiple model types. These are then combined in an ensemble-learning approach to enhance predictive modelling. This approach will be used to model both Alzheimer’s subtypes and the stages of Alzheimer’s and cognitive decline previously analysed in the ADNI datasets. In addition to this hypothesis-free, data-driven features selection approach, genomic signals from circadian and sleep-associated loci will be independently investigated and associated with other features in a planned sub-analysis. Anticipated Results: Our anticipated results will follow four strands of the analysis described above. Using patient-level information presented in clinical and psychiatric assessments available in the ADNI resource, we will project patient characteristics into the phenotype space by integrating features from several well-developed biomedical ontologies, potentially creating a new depiction of Alzheimer’s symptoms. Importantly, profiles of subjects with similar clinical features will be created, leading to patient subgroups for further structural and molecular characterization. Genomic analyses of these cohorts will leverage tests for epistatic interactions between genes, which will create gene-level networks which help characterize ADNI subpopulations. Features derived from structural MRI will likewise be created, and features which discriminate between phenotype-driven subtypes will be investigated. The combination of genomic and imaging features for each subtype of ADNI subjects identified will be tested together, and their ability to classify subjects into correct cohorts will be thoroughly evaluated. Lastly, correlations between features themselves will be investigated, focusing on circadian-related genomic loci and their relationship to structural morphology in ADNI subjects. This will, ideally, evaluate the relationship between loci associated with abnormal circadian biology, cognitive decline, and the structural brain biology of ADNI subpopulations. Conclusions: Our combination of patient-subgroup gene association with brain region-specific structural study will provide a novel set of potential biomarkers (features) to subset a population with a diverse phenotypic profile. While basic aims in this proposed work include method patient characterization and multiple machine learning approaches applied to many datasets, the ultimate aim is to improve knowledge of which genetic lesions and brain morphological characteristics contributing to Alzheimer’s patient phenotypes. References: 1. Hoehndorf R, Schofield PN, Gkoutos GV. PhenomeNET: a whole-phenome approach to disease gene discovery. Nucleic Acids Res. 2011 Oct 1;39(18):e119–e119. 2. Boudellioua I, Mahamad Razali RB, Kulmanov M, Hashish Y, Bajic VB, Goncalves-Serra E, et al. Semantic prioritization of novel causative genomic variants. PLoS Comput Biol. 2017 Apr;13(4):e1005500. 3. Gkoutos GV, Schofield PN, Hoehndorf R. The neurobehavior ontology: an ontology for annotation and integration of behavior and behavioral phenotypes. Int Rev Neurobiol. 2012;103:69–87. 4. Hoehndorf R, Hancock JM, Hardy NW, Mallon A-M, Schofield PN, Gkoutos GV. Analyzing gene expression data in mice with the Neuro Behavior Ontology. Mammalian Genome. 2014 Feb;25(1–2):32–40. 5. Henry V, Moszer I, Dameron O, Potier M-C, Hofmann-Apitius M, Colliot O. Converting Alzheimer s disease map into a heavyweight ontology: a formal network to integrate data. arXiv:180710509 [q-bio] [Internet]. 2018 Jul 27 [cited 2019 Dec 13]; Available from: http://arxiv.org/abs/1807.10509 6. Malhotra A, Younesi E, Gündel M, Müller B, Heneka MT, Hofmann-Apitius M. ADO: a disease ontology representing the domain knowledge specific to Alzheimer’s disease. Alzheimers Dement. 2014 Mar;10(2):238–46. 7. Köhler S, Carmody L, Vasilevsky N, Jacobsen JOB, Danis D, Gourdine J-P, et al. Expansion of the Human Phenotype Ontology (HPO) knowledge base and resources. Nucleic Acids Res. 2019 Jan 8;47(D1):D1018–27. 8. Pesquita C. Semantic Similarity in the Gene Ontology. In: Dessimoz C, Škunca N, editors. The Gene Ontology Handbook [Internet]. New York, NY: Springer; 2017 [cited 2019 Dec 20]. p. 161–73. (Methods in Molecular Biology). Available from: https://doi.org/10.1007/978-1-4939-3743-1_12 9. Rathore S, Habes M, Iftikhar MA, Shacklett A, Davatzikos C. A review on neuroimaging-based classification studies and associated feature extraction methods for Alzheimer’s disease and its prodromal stages. NeuroImage. 2017 Jul 15;155:530–48. 10. Chincarini A, Bosco P, Calvini P, Gemme G, Esposito M, Olivieri C, et al. Local MRI analysis approach in the diagnosis of early and prodromal Alzheimer’s disease. NeuroImage. 2011 Sep 15;58(2):469–80. 11. Boyle EA, Li YI, Pritchard JK. An Expanded View of Complex Traits: From Polygenic to Omnigenic. Cell. 2017 Jun 15;169(7):1177–86. 12. Wu X, Dong H, Luo L, Zhu Y, Peng G, Reveille JD, et al. A Novel Statistic for Genome-Wide Interaction Analysis. PLOS Genetics. 2010 Sep 23;6(9):e1001131. 13. Ueki M, Cordell HJ. Improved Statistics for Genome-Wide Interaction Analysis. PLOS Genetics. 2012 Apr 5;8(4):e1002625. 14. Basu S, Kumbier K, Brown JB, Yu B. Iterative random forests to discover predictive and stable high-order interactions. PNAS. 2018 Feb 20;115(8):1943–8. 15. Kumbier K, Basu S, Brown JB, Celniker S, Yu B. Refining interaction search through signed iterative Random Forests. arXiv:181007287 [cs, stat] [Internet]. 2018 Oct 16 [cited 2019 May 10]; Available from: http://arxiv.org/abs/1810.07287 16. Bravo-Merodio L, Williams JA, Gkoutos GV, Acharjee A. -Omics biomarker identification pipeline for translational medicine. J Transl Med. 2019 May 14;17(1):155. |
