In many population genetic models, mutations are assigned a fitness effect - an assigned genotype -> fitness map.  Simulations can then explore how gene frequencies change over time under a number of demographic scenarios.  An alternative is to use a generative model - often times a set of differential equations - that use genotypic information to produce a phenotype and then map this phenotype to a fitness value.  We are using such a simulation framework to study how genetic networks evolve under different demographic situations.  Experience with C++ would be very useful for this project since the basic population genetic framework consists of a library of C++ templates.

C. elegans embryos are powerful model systems for studying developmental variation using microscopy because they are transparent and can be fluorescently marked. However, automatic detection/segmentation of fluorescently marked nuclei is a challenging image informatics problem. This rotation would consist of applying deep learning techniques to this problem based on a corups of images that have been previously manually curated. Experience with python is useful and previous exposure/experience with deep learning would also be useful.

The Palmer lab hosts a NIDA-funded National Center of Excellence to examine the role of genetic differences in a variety of complex rat behaviors related to drug abuse (www.ratgenes.org). As of March 2020, we have already phenotyped and genotyped over ~7,000  rats as part of this project, and have secured funding to collect an additional 8,000 rats. In addition, we have previously collected similar data from more than 4,000 rats from a different population as well as several thousand mice and are now in the process of collecting similar data on over 5,000 zebra fish. In addition to genotypes at millions of SNP markers, and complex behavioral observations spanning weeks to months, we are also examining RNASeq (including single cell RNASeq), Epigenetic, microbiome and metaboloic data from these populations. The analysis, integration and development of new methods for analysis create numerous opportunities for bioinformatics projects and interactions with the bench scientists who are creating these data. We also collecting human genetic data and are exploring various methods for integrating human and non-human datasets.