Hi all,

I am looking for two postdocs to work on developing and applying methods for analysis of GWAS data, based at Aarhus University, Denmark. Applications by 1st September. A summary of the positions is below, or for more details see this link. If you have any questions, please email me at [log in to unmask]

https://www.nature.com/naturecareers/job/two-postdocs-developing-gwas-methods-with-a-focus-on-prediction-of-complex-traits-aarhus-university-au-728333

Thanks, Doug

--------------------------------------------------

Aarhus University is recruiting two 2-year postdocs in statistical genetics based at the Center for Quantitative Genetics and Genomics (QGG). The provisional starting date is 1st November 2020.

The positions
The primary supervisor will be myself (Doug Speed, QGG), with co-supervision from Drs Bjarni J Vilhjalmsson (NCRR) and Søren Østergaard (Department of Clinical Medicine). Below are the two main aims of the positions, however the specific projects will be decided according to the interests and experience of the successful applicants.

1 – Heritability analysis of genome-wide association study (GWAS) data.
I have created the software package LDAK which contains a variety of methods for analysing GWAS data. These include tools for estimating SNP heritability, heritability enrichments and genetic correlations and for performing single-SNP and gene-based association tests. Each of these methods starts with the linear mixed model. Possible projects include extending these methods or developing new versions (e.g. adapt them to accommodate different datatypes, generalizing them to multiple phenotypes, or create versions that require only summary statistics). The primary focus will be methods for human data (e.g., UK Biobank), but there is also the opportunity to work with animals and plants.

2 – Using artificial intelligence to improve genetic prediction of complex traits
This work will be part of a project investigating whether artificial intelligence methods can be applied to genetic data in order to improve prediction of complex traits. As described in the project description “Many complex diseases are highly heritable (e.g., schizophrenia, major depression, Ischemic Stroke and Alzheimer’s Disease all have heritability between 40 and 80%), and so it should be possible to accurately predict which individuals will develop them based on genetic information. However, at present this is not the case. For most complex diseases, the best prediction models have accuracy less than a fifth the theoretical maximum. In an attempt to move the field of personalized medicine forward, we will adapt tools for natural language processing (NLP), a branch of artificial intelligence, for use with genetic data.”

Relevant references
1<https://www.nature.com/naturecareers/job/two-postdocs-developing-gwas-methods-with-a-focus-on-prediction-of-complex-traits-aarhus-university-au-728333#fn1> Speed et al. Reevaluation of SNP heritability in complex human traits (Nature Genetics, 2017)
2<https://www.nature.com/naturecareers/job/two-postdocs-developing-gwas-methods-with-a-focus-on-prediction-of-complex-traits-aarhus-university-au-728333#fn2> Speed and Balding. SumHer better estimates the SNP heritability of complex traits (Nature Genetics, 2019)
3<https://www.nature.com/naturecareers/job/two-postdocs-developing-gwas-methods-with-a-focus-on-prediction-of-complex-traits-aarhus-university-au-728333#fn3> Speed et al. Evaluating and improving heritability models using summary statistics (Nature Genetics, 2020)

Supervisors and supervision
I specialize in developing statistical methods for analyzing large scale GWAS data. I have released the software LDAK which contains tools for detecting causal variants, constructing prediction models and better understanding genetic architecture, using both individual-level data and summary statistics (see www.ldak.org<http://www.ldak.org/> for more details).

I believe that when performing a statistical analysis, it is very important to understand what the analysis is doing. Further, if you understand an analysis, it increases the chance that you can find ways to improve the analysis or to transfer the ideas to other problems. In general, I only use software that I could in theory code up myself (I say in theory, because it would be very inefficient to always make my own software). Therefore, I am keen that people I supervise also understand the analyses they perform, and am happy if they spend time trying to understand methods (I will also try and help explain methods, where I can).

Bjarni J Vilhjalmsson is interested in developing and applying statistical methods that integrate health records and large genetic datasets to study the etiology of diseases and psychiatric disorders. This includes causal inference, heritability analysis, GWAS, as well as polygenic risk scores.

Søren Dinesen Østergaard is a medical doctor who focuses on psychiatric research. He is particularly interested in translational psychiatry, and the idea that studies should cover the full pathway from discovery in the lab, bench to bedside, bedside to clinical applications, and from clinical applications to healthcare and global health.

Qualification Requirements
A PhD degree and strong expertise in statistical genetics is essential. The positions will involve analysis of large-scale genetic datasets, so ideally candidates would be familiar with popular genetic software (e.g, PLINK) and at least one coding languages (e.g. R). When applying, please provide a cover letter that details which statistical softwares or GWAS methods you familar with. These can include GWAS quality control, linear regression, logistic regression, survival analysis, linear mixed models, Mendelian Randomization, polygenic risk scores, LDSC, LDPred, REML, BLUP, Haseman Elston, Bolt-LMM, Random Forests, NLP, BERT, or any others that might be in any way relevant to the position. For each software/method, please indicate whether your experience is standard (i.e., able to run an analysis using existing software) or expert (i.e., able to code your own version).

Center for Quantitative Genetics and Genomics
The QGG is a major center for research and education in quantitative genetics and quantitative genomics (ww.qgg.au.dk/en). It has about 70 employees and visiting researchers from over 15 countries. It is world-famous for its research in both animal and plant breeding, however, it is now wishes to focus more on human traits. Researchers at QGG work in teams, and the center strives for an excellent working environment. It therefore values self-leadership, open communication, respect for diversity and success in mentoring, supervision and teaching. The QGG is in the process of moving. Therefore, the successful applicants can choose between working in Foulum (the current location of QGG) or Aarhus Centre (its location from late 2021).


You may leave the list at any time by sending the command

SIGNOFF allstat

to [log in to unmask], leaving the subject line blank.