Biological Discovery and Consumer Genomics Databases Activate Latent Privacy Risk in...
summary
Claims that summary omics (gene expression, methylation) can be linked to unique genomes by inferring genotype from QTL signals and then matched to GEDmatch or similar consumer genealogy databases. Methods reconstruct enough eQTL-inferred genotypes to identify relatives out to first/second degree in many cases; prototypes upload simulated profiles, demonstrate real-world re-identification (examples include a diabetes study outing a pilot) and argue public omics become persistent time bombs as tools and databases improve.
AI-written summary · about
chapters (11)
- 0:00Intro
- 0:32Golden State Killer Sentenced to Life in Prison Without Parole
- 4:59How can personal privacy be breached by biological data? (Linking attack)
- 6:43Omics reads are shared with controlled access, but summary data have unrestricted sharing
- 7:03Genotype can be inferred from gene expression via expression quantitative trait loci (eQTLs)
- 8:29Gene expression linking was deemed a theoretical risk, but not a practical one
- 9:14The ability to link gene expression to genomic data increases over time
- 11:38Is there a real-world risk?
- 13:56Privacy risks emerge in research data
- 15:28Many types of functional genomics data also have privacy risks
- 17:41Security is expected to degrade over time
related talks
01h 10mENCODING IDENTITIES: THE CASE OF COMMERCIAL DNA DATABASE
01h 11mGlobal Perspectives of Regulating Genetic Data in Scientific Research: Beyond Privacy?
01h 04mPrivacy Dependencies
08mAnne Wojcicki interviewed at Web 2.0 Summit 2011
16mThe Secret Life of Data - Part 3
36mOn Beyond Genome: Opportunities and Challenges Using the Other Omics for Precision Medicine
10mHealth Data Anonymization and Privacy
29mPrivacy in Epigenetics: Temporal Linkability of MicroRNA Expression Profiles