Our research projects cover a wide spectrum of computational biology, software engineering, and biomedical data science applications with an emphasis on translational and clinical applications. We develop methods and apply them to uncover the molecular mechanisms of disease and to identify the molecular alterations that cause disease. Projects span diagnosis, prognosis and therapeutic response prediction in both rare diseases and cancer.
In addition to many ongoing large scale omics data analysis projects, molecular diagnostics programs, and case studies, we actively develops methods, algorithms and tools to help others perform these same types of analyses.
We have published articles and case reports with major areas of focus in precision medicine, genomics, immunogenomics, biomarker discovery, causal variant identification, sequence analysis, gene regulation, and proteomics. We are also actively involved in training the next generation of computational biologists through formal courses, mentoring, tutorials, and workshops.
Projects can be artificially divided into those that are primarily specific disease focused and those which are primarily methods and tool development. There is a lot of crossover of course between them with novel methods being developed to address the needs of a particular biomedical project.
Primarily disease focused projects
Rare, misdiagnosed, or undiagnosed disease programs
We work with clinicians from many institutions and specialties to provide definitive molecular diagnoses. Patients or parents of individuals with genetic disorders and those who take care of them generally have a number of questions they would like answers to. These include: What is the diagnosis? How did it happen? Who else is at risk? What can be done? What can we expect? In CGDS we want to help answer these questions by extracting knowledge from omic and associated data using tools and methods we apply and/or develop. We are interested in applying our methods widely and have established collaborations with clinicians and researchers spanning a number of rare or not so rare diseases.
Pediatric cancer cardiotoxicity
Survivors of pediatric cancers are at high risk of developing heart disease both during treatment and after their cancer is defeated. Working in collaboration with Dr. Smita Bhatia’s UAB Institute for Cancer Outcomes and Survivorship research group, we are working to uncover causes of cardiotoxic consequences in these patients. This involves case by case analysis to identify rare large effect variation as well as polygenic risk analyses to identify associations with smaller impact more common variation.
Cystic Fibrosis (CF)
Cystic Fibrosis is a rare disease that displays heterogeneous symptoms and a non responder rate of 30%. Patients within the same family, who have the same molecular cause of CF, can have very different symptoms, progression, and outcome. CGDS works with Dr. Phil Farrell and Dr. HuiChuan Lai at UW-Madison and other members of the CF-FIRST consortia to identify genetic modifiers in patients with CF that are related to their symptoms and response to therapeutics making use of candidate variant and polygenic risk analyses.
Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS)
ME/CFS is a complex disorder leading to exertional intolerance, chronic fatigue not helped by rest, pain syndromes, as well as a multitude of other multi-system symptoms. Working with Dr. Jarred Younger’s group at UAB, we are analyzing patient medical histories, as well as WGS and RNA-seq data to identify molecular variation associated with onset or symptoms of ME/CFS.
Ehlers Danlos Type III (EDS3)
EDS3 is a connective tissue disease primarily characterized by the hypermobile joints which can lead to spontaneous dislocations and pain. Other multi-system symptoms are also commonly seen. For many patients, the genetic cause of this disease has not been identified. We are working with Dr. Daniel Greenspan at UW-Madison to undertake analysis of a kindred with EDS3 to attempt to find the causal variant.
Muscular Dystrophies (MD)
Several distinct forms of muscular dystrophy exist. In all molecular variation leads to muscle degeneration and most individuals with this condition will eventually need a wheelchair. Other symptoms include trouble breathing or swallowing. Some patient also experience neurologic symptoms including autism like behaviours. Working in collaboration with UAB’s Drs. Matt Alexander and Mike Lopez and their collaborator Dr. Luca Bello we are undertaking studies to identify causal variation in patients where existing tests were not diagnostic as well as studies aimed at identifying modifier variation.
Prader-Willi Syndrome (PWS)
Molecular variation in the genetic imprinting region on Chr 15 cause PWS, which presents with a wide variety of symptoms including behavioral problems, intellectual disability, and short stature. There is a wide range of symptoms and severity in individual patients and they are not always tied to the molecular variation present in the individual. We are collaborating with Dr. Theresa Strong of the Foundation for Prader Willi research to generate a comprehensive data repository to begin to unravel these complex genotype-phenotype associations.
Ciliopathies are human disorders that arise from the dysfunction of motile and/or non-motile cilia. At least 35 different ciliopathies collectively affect nearly all organ systems, with prevalent phenotypes including polycystic kidney disease, retinal degeneration, obesity, skeletal malformations and brain anomalies. We are working with Drs. Brad Yoder, Michal Mrug and others at UAB to understand the molecular underpinnings of these diseases and to prioritise variants of uncertain significance for functional confirmation in a variety of animal models.
Software development focused projects
In order to tackle these projects at times we need to develop tools where none exists or where the cost of commercial tools would inhibit progress. We are interesting in a variety of projects aimed at extracting knowledge from omic and associated data; to be successful we have developed a number of tools or pipelines.
A major focus of the team is the development and application of tools to support definitive molecular diagnosis and increased understanding of variant impact. Delphi is our UAB wide precision medicine platform integrating biomedical, computational biology, and data science approaches as well as state of the art software development methods to create a fast and effective tool for variant annotation, analysis, prioritization, interpretation, and reporting. Building on more than a decade of experience in developing tools, Delphi improves multi-omics analysis in identification and interpretation of causal and modifier molecular variation.
Our Apollo dashboard was developed to support our colleagues in the UAB pathology and genetics molecular diagnostic labs by aiding lab members and directors track the necessary components for COVID testing.
CGDS always works towards improving patient’s lives today, but also looks to the future of science and medicine and the tools needed for tomorrow. Ditto is a tool that makes use of scalable machine learning methods to aid in more efficient and timely molecular diagnosis as well as uncovering patterns of variation and effect that may support identification of specific therapies for specific patients.
CGDS works with a variety of collaborators who work on generation and testing of animal models of genetic diseases to study cellular and biochemical impact of variation. Our variant analysis platform Delphi was initially developed for variant analysis, interpretation, and reporting in a molecular diagnostic setting. The Delphi platform will now be built out to support multi-omics functional testing analysis for prioritization of human variation for functional testing.
Secondary analysis tools
In addition to diagnosing disease, performing analysis and developing tools CGDS undertakes a variety of DevOps related tasks to meet the needs of doctors and researchers at UAB. We do this by utilizing cutting edge technology and best software engineering practices to develop efficient and effective tools and processes that make possible new developments and discoveries in medicine and clinical information, data science, bioinformatics and analysis. We develop a variety of pipelines and packages to support data generation, analysis, exploration, visualization, and interpretation. Where possible we implement or collaborate to implement existing tools, but we also develop pipelines and methods where necessary. We have made a number of these available and all pipelines in the future will be widely shared.