Applied statistical methods and network analysis for identifying similarities and differences in biological data for various traits.
Abstract
"Observing differences in various traits, such as differences in human height, why some people are more susceptible to certain diseases, or why some species survive in certain environments, we often question why and how these differences appear. Studies of genetic differences and biological compounds’ effect on the traits can often give some answers to these questions. With the high-throughput techniques available today, scientists have the opportunity to analyse a broad variety of data, from genetic sequences, to specific compounds’ effects on biological pathways, and further, their effects on an organism’s traits. Several studies have been published with success stories of how analyses of certain types of data can provide new insights. However, much is still unknown, and combining these analyses could contribute to further insights than if the data were analyzed separately.
In this PhD-work, the overall aim is to combine existing methods for analyzing genetic and biological differences in various traits, and to obtain new insight into how these genetic and biological differences and their interplay affect traits. We combine methods from three different fields, statistics, network analysis, and genetic association studies, to gain new insight into our chosen traits. In Paper 1, we studied differences between four populations of stickleback fish, using statistical shrinkage and regression models together with differential co-expression networks of measured protein abundances, to explore the genetics of how these populations could inhabit highly diverse environments. In Paper 2, we combined Genome-Wide Association Studies (GWAS) with post-GWAS analyses and gene co-expression network analysis to study genetic differences between patients experiencing recurrent events from single events of two prevalent and severe diseases, atrial fibrillation and myocardial infarction. In Paper 3, we used summary statistics obtained from Phenome-Wide Association Studies (GWAS applied to several phenotypes/diseases) to create disease networks where diseases were linked through common genomic associations. By further comparing with observed co-occurrences, we analyzed the genetic and observed co-occurrence associations between diseases at a system level.
Taken together, this Phd-work highlights the value of combining methods from different fields of science, and it demonstrates how such combinations may provide further insight into genetic and biological compounds’ effect and interplay in diseases and traits. With our analysis, we do identify genetic differences between the studied traits, we identify genes and gene interactions specific for the traits, and we generate hypotheses for genetic effects that need further studies. Finally, we have identified and studied groups of genetically linked diseases. Hopefully, this work will provide a lasting contribution to the studies of genetic and biological differences and similarities and their effect on diseases and traits."
Has parts
Paper 1: Hall, Martina; Kültz, Dietmar; Almaas, Eivind. Identification of key proteins involved in stickleback environmental adaptation with system-level analysis. Physiological Genomics 2020 ;Volum 52.(11) s. 531-548. Copyright © 2020 the American Physiological Society. This Paper is not included due to copyright restrictions. Available at: http://dx.doi.org/10.1152/physiolgenomics.00078.2020Paper 2: Hall, Martina; Skogholt, Anne Heidi; Surakka, Ida; Dalen, Håvard; Almaas, Eivind. Genome-wide association studies reveal differences in genetic susceptibility between single events vs. recurrent events of atrial fibrillation and myocardial infarction: the HUNT study. Frontiers in Cardiovascular Medicine 2024 ;Volum 11. s. – Published by Frontiers Media. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). Available at: http://dx.doi.org/10.3389/fcvm.2024.1372107
Paper 3: Hall, Martina; Skinderhaug, Marit K.; Almaas, Eivind. Phenome-wide association network demonstrates close connection with individual disease trajectories from the HUNT study. Accepted for publication in PLOS ONE.