Week 3 QTO Mapping

QTL Mapping Overview of Methods Basic Elements of Mapping QTL 3 sources of information we need for QTL analysis: Phenotype of trait: Score what phenotype is in the individuals Marker Genotype: Score marker genotypes for molecular markers identified throughout the genome (SNPs for example) Population: What is the genetic structure of the mapping population? Methods to Genetically Dissect Complex Traits 1) Linkage Analysis (in pedigrees and experimental crosses) 2) GWAS (human, plant and animal populations) 3) Allele-Sharing Methods (in pedigrees) Linkage Analysis ➢ Can figure out which regions of the genome control variation of a trait ➢ Two varieties of plants for example can be illustrated in different colours (green and blue on the graph) ➢ Using controlled crosses, we create segregating population (F2 generation) with a mosaic of chromosomal chunks from the parents. ➢ Within in this population, we score individuals for their genotype and phenotypes. ➢ Can determine which segments determine trait variation in linkage analysis. Association Analysis: GWAS ➢ Uses naturally occurring populations of individuals (difference to linkage analysis which uses F2 generation) ➢ Pulls out peaks / dots on graph that indicate a QTL ➢ Shows many generations in which chromosomes have been re-shuffled and new genetic variation combos have been made ➢ You can localise QTLs to much more narrow windows Statistical Concepts Calculating a Likelihood Ratio A likelihood ratio is the ‘ratio of the two probabilities. → Prob 2 / Prob 1 = 1403 = Likelihood Ratio / Odds ratio It is ‘1403’ more times likely that probability 2 is correct / more likely than probability 1. Significance expressed as a ‘LOD’ score (Log of Odds) The next step is to use this likelihood ratio and express it as a LOD score. This ratio has a complex distribution, and we can turn it into an easier format to work with. Any LOD > 3 is considered ‘significant’ because the P value is at least less than 1/1000. LOD = log10 (L2 / L1) LOD = log10 (1403) = 3.15 More Likelihood Ratios… We can calculate the likelihood of a QTL in any particular position in the genome over likelihood of no QTL in that position. Likelihood = Probability (Data I QTL) / Probability (Data I No QTL) Once we obtain this likelihood ratio, we may convert it into ‘LOD’ score and apply the threshold to determine probability significance. At a particular position on the chromosome, there is an odds ratio of 100 that there is a QTL. What is the corresponding LOD score? Answer: Log 10 (100) = 2 There is no significant QTL as LOD score is < less than 3. Linkage Mapping Method Principles of Single Marker Analysis (SMA) SMA in an F2 population For the QTL allele: we are not directly able to observe the genotype at the QTL locations because we do not know where the QTL are. We use markers to infer whether or not a QTL is present (by genotyping the markers). We model the genotypes for quantitative traits. Linear model of jth F@ individuals phenotype value, Yj Defining a Genetic Effect for the 3 Major QTL Genotypes When the marker is unlinked to the QTL (r =1/2) When the Marker is linked to the QTL (r < ½) ➢ Mean of phenotype in each group should be different and we can observe which marker has a QTL nearby ➢ Can estimate size of the genetic effect of the QTL, including additive and dominance effects ➢ Cannot however distinguish large QTL effect size from close linkage (small r) Example Question: Single Marker Analysis Tomato plants from a backcross were scored for fruit weight and genotyped as B/B or B/S allele for 5 markers and the frequency distributions are shown: Which of the markers, if any, show evidence of a nearby QTL? For markers ‘M1’ and ‘M5’, the distribution of the phenotype score is exactly the same despite the marker genotype. There is not a likely QTL in these positions M3 shows a shift of phenotype distribution to the right (BB have higher fruit weight than BS). This indicates that somewhere near M3, there must be a QTL present as the two distributions look different. The QTL may be between M2 and M3. The QTL is likely to be closer to M3 than M2, as there is a bigger phenotypic change in the distribution of M3. Which allele is “A+” i.e., increases fruit weight? Looking at M3, we can judge that the BB allele is responsible for increasing fruit weight as there is a shift of the phenotype distribution to the right, compared to a shift to the left with BS allele. Pairwise Marker Analysis: Interval Mapping ➢ We plot the relative belief of the presence vs absence of QTL over every possible genome position (in between and at markers) ➢ QTL are located where LOD is > than 3. Any parts of the lines that go above the threshold (3) are considered QTL. In the genome above, the threshold seems to be crossed in chromosomes 1, 3, 4, 5 and 7. Methods of Interval Mapping Test every possible position of the chromsoome to see if there is any evidence for a QTL being present. Principles of Interval Mapping How can we use marker information to infer whether a QTL may be present at that position? R1 = recombination frequency between marker 1 and QTL R2 = recombination frequency between QTL and marker 2 Based on genotype at the flanking markers, we work out probability of each QTL genotype. Depends on occurrence of recombination between markers and QTL. In the above example, there is only recombination in ‘interval 4’ as there is a capital ‘Q’ and small ‘m2’ genotype.