NEW YORK – Gretel, a data privacy firm, on Tuesday announced a partnership with Illumina to generate artificial genomic data for use in medical research.
According to Gretel, the firms will provide researchers with statistically accurate, artificial versions of genomic datasets that align with GDPR, CCPA, and other privacy laws, the use of which will allow them to avoid lengthy approval processes by institutional review boards.
Financial and other details of the partnership were not disclosed.
Synthetic data involves the use of artificial intelligence algorithms "to create artificial versions of sensitive datasets," the firms wrote in a white paper released on Tuesday.
For a case study described in the white paper, they synthesized a mouse genotype dataset that provided similar results to the real dataset in a genome-wide association study analysis. The compute cost for training synthetic models on the genotype and phenotype data was approximately $1,440.
The researchers claimed this study demonstrates the viability of new use cases for synthetic data in genomics, diagnostics, forensic biology, virology, and other related fields.