Genetic relationship lab answers

  1. [Coefficient of coancestry] Why do we see blocks of patterns in the heatmap? Is there something you can infer from the pedigree?
    • Each individual in the third generation have one full-sib.
  2. [Coefficient of fraternity] Can you describe any difference between the heatmaps for coancestry versus fraternity?
    • Coefficients for coancestry are higher than fraternity, because there is not much inbreeding within the population.
  3. [Parentage inference] Can you roughly count the number of recombination events in one RIL?
    • RIL01=34, RIL02=25, RIL03=34, RIL04=28, RIL05=27, RIL06=25, RIL07=28, RIL08=32, RIL09=20, RIL10=28.
  4. [Additive GRM] Why are the values in the diagonals larger than the values in the off-diagonals?
    • The diagonals are the coefficients between an individual and itself so their P(IBD) are higher.
  5. [Dominance GRM] Do we need D-GRM in an inbred population, and why?
    • No, because all of the off-diagonal values would be 0.
  6. [Compared A/D-GRM to coancestry/fraternity] Why is there such a large discrepancy between the estimates from pedigree- vs marker-based coefficients?
    • The pedigree estimates are based on the expected probability for the given pedigree. The marker-based coefficients are closer to reality as genetic inheritance is mostly random.
  7. [Additive x additive GRM] Can you explain why the AA versus A values have a quadratic/U-shaped distribution?
    • The AA coefficients are computed as the A coefficients squared, hence the quadratic pattern.
  8. [Splitting and combining GRM] Why is the LOCO-GRM not perfectly correlated to the original GRM?
    • The LOCO-GRM is missing information from one chromosome.
  9. [Fst] Which (sub)-population is most genetically distant from the overall population, and why?
    • Teosinte, because it has the highest Fst value of all.
  10. [Genome-wide Fst] If 0.15 is out cutoff, do we have any significant marker differentiating the two (sub)-populations?
    • Yes, on chromosome 2, 7 and 9.
  11. [Roger’s/Nei’s distance] Can you describe any difference or similarity between these two trees?
    • The clusters around B73 and Mo17 are quite similar between the two trees, but the overall trees are quite different.
  12. [PCA] Roughly how many principal components are needed to explain 50% of the variation seen here?
  13. [PCA] Can you describe what are each of the principal components separating?
    • PC1=Within PVP, PC2=Between PVP-CIMMYT-Teosinte-EU_Inbred, PC3=Within PVP, PC4=Between PVP-CIMMYT/Teosinte.
  14. [Population admixture] Which K do you think is the right one for this dataset?
    • 2, W22 and FV2 are the only two founders.
  15. [GWAS] Can you describe what is the difference in the GWAS results with and without population structure control?
    • There is a false positive on Chromosome 6.
Updated on February 5, 2024