Yi Ding (@yi_ding_) #ASHG22: Polygenic scoring estimates genetic value by aggregating signal across many loci. Potential for polygenic scores has already been well established -- but so have the challenges, especially the low accuracy in non-European populations.
Yi Ding (@yi_ding_) #ASHG22: Genetic ancestries are continuous and cluster-free.
- By forcing clusters on ancestry, miss inter-individual variability.
- The boundaries between these groups is also fairly arbitrary.
- Clustering also typically removes any unclassified individuals
- By forcing clusters on ancestry, miss inter-individual variability.
- The boundaries between these groups is also fairly arbitrary.
- Clustering also typically removes any unclassified individuals
Yi Ding (@yi_ding_) #ASHG22: Goal is to evaluate PGS accuracy at an individual level.
Using genetic distance to characterize and individual's unique position (Euclidiean distance) in genetic ancestry continuum. Also evaluating individual PGS accuracy.
Using genetic distance to characterize and individual's unique position (Euclidiean distance) in genetic ancestry continuum. Also evaluating individual PGS accuracy.
Yi Ding (@yi_ding_) #ASHG22: As genetic distance decays, so does the PGS accuracy in the UKBB data. See the same trend in the UCLA ATLAS dataset.
Still see this *within* "homogenous" genetic ancestry clusters (e.g. Europeans). Even more obvious in admixed populations.
Still see this *within* "homogenous" genetic ancestry clusters (e.g. Europeans). Even more obvious in admixed populations.
Yi Ding (@yi_ding_) #ASHG22: Average correlation of genetic distance and PGS accuracy is -0.98 for UKBB and -0.95 for ATLAS (still -0.62 for Europeans in ATLAS).
• • •
Missing some Tweet in this thread? You can try to
force a refresh
