In biomedical research, it is common to perform numerous analyses of genomic data, for example, to understand the cause of a particular disease. Regulatory laws protect the privacy of individuals but hinder access to genomic data. One solution to this is the development of bioinformatic tools to create synthetic DNA data. One of the challenges is to capture genomic diversity representative of differences within and between populations, especially for rare genetic diseases. In this study, we present SynthDNASim, a tool for creating diverse synthetic DNA datasets. Our approach is to create diverse DNA datasets taking into account factors of genetic evolution and ancestry with Huntington's disease (HD) as a use case. In particular, with HD variants from European, African, and Middle Eastern populations. We will show our tool and future plans on applying semantic methods and tools to make SynthDNASim more FAIR (Findable Accessible Interoperable Reusable).
In biomedical research, it is common to perform numerous analyses of genomic data, for example, to
understand the cause of a particular disease, or genetic processes, or to identify gene variants. One of
the difficulties in these analyses is the collection, storage, use, and reuse of genomic data because an
individual's genomic data is private. Especially if an individual has a rare disease like Huntington’s
disease (HD) it is theoretically possible to retrace the DNA to this individual. Thus, there are
regulatory laws that protect the privacy of these individuals but hinder access to genomic data. One
solution to this is the development of bioinformatic tools to create diverse synthetic DNA data so that
researchers in biomedical research can create synthetic DNA data and make the research faster and
more reproducible. [
________________________
In Figure 1 we illustrate the SynthDNASim pipeline. The first step is the retrieval and pre-processing of genomic information from different data sources: National Center for Biotechnology Information (NCBI) for the SNP variants, NCBI for the sequence of Chromosome 4, and lastly the user input. Next is a sequence of steps to create the synthetic DNA sequences per population. Python is used for the user input, creating the config file (JSON file). Each sequence has its own metadata including haplotype, genetic variants, CAG repeats, gene, chromosome, etc.
The remaining work of this project is to perform a validation on the generated data and to use semantic
methods and tools to make the project more FAIR. One option for this is to create the metadata for the
output data and the tool. For the creation of the output metadata Data Catalog Vocabulary (DCAT) can
be used. [
We want to thank Alex Stikkelman and the 4MedBox team for all their support and help. We also want to thank Ivo Fokkema and the Biosemantics group at the LUMC for their input and help in this project. This project received funding from 4MedBox. N. Queralt-Rosinach is supported by funding from the European Union’s Horizon 2020 research and innovation program under the EJP RD COFUND-EJP N° 825575 and by a grant from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 847826 (Brain Involvement iN Dystrophinopathies (BIND)). We would like to thank to the EJP RD and BIND for supporting research on generating synthetic health data for rare diseases research.