Maciej Trzaskowski

My PhD focused on addressing the problem of ‘missing heritability’; this refers to the discrepancy between the high heritability estimates for many complex traits derived from twin studies, and the very small amount of variance explained from known common genetic variants identified through genome-wide association studies (GWAS). Researchers proposed several hypotheses to explain this phenomenon. However, I was the first to directly address this issue by bringing quantitative and molecular genetic methods together. I applied structural equation twin models, GWA, genome-wide complex trait analysis (GCTA) and genetic risk scores comprising multiple single nucleotide polymorphisms (SNPs) to a longitudinal sample of the same individuals to control for sampling, measurement and change over time. This unique design allowed me to directly evaluate the extent of the missing heritability. Using cognitive development, behavioural problems and body mass index (BMI), I showed that the additive genetic effects of all common SNPs (SNP heritability) can account for some of the missing heritability, but a nontrivial proportion of it is likely to be caused by influences such as ‘uncaptured’ rare variants (which whole-genome sequencing will address), non-additive genetic influences (e.g., gene-gene interactions, dominance), interplay between genes and environments, or some combination of them. My PhD thesis comprised six first-author peer-reviewed publications.   

My work over the last four years extended beyond the PhD requirements and resulted in total, in 19 peer-reviewed papers of which I was first author on nine. During this time I have learnt and applied an extensive selection of statistical methods for both quantitative and molecular genetics. In particular, I have applied a wide variety of structural equation models to twin data, translated many of them from Python to R, run GWAS, created genetic risk scores for complex traits comprising multiple SNPs, and conducted GCTA. I have become proficient in a number of programming languages to automate processes and facilitate more complex analyses (e.g. R, Unix, Shell Scripts, Java), and I explored many bioinformatics tools, such as genome browsers, SNAP, Galaxy, etc. Developing this suite of skills has enhanced my capacity to develop new methods. I have a strong methodological propensity and am highly motivated to explore new methods.