Modeling Genome-wide by Environment Interactions through Omnigenic Interactome Networks by Wang et al. 2021, Cell Reports

Published: 9 April 2021| Version 1 | DOI: 10.17632/9nrsp5wfvz.1
Rongling Wu


Phenotyping Imaging: Experiment of Euphrates poplar rooting capacity in transparent tubes via tissue culture. Root images were monitored repeatedly once every 5 days until the 78th day when growing roots fill tubes. Related to Figure 1. Go Analysis: Chromosomal position, marker type, allele type, and functional description of each QTL detected by biFunMap for the growth of lateral roots in the F1 mapping population of Euphrates poplar. Related to Figure 1. Functional Mapping: Simulation result about functional mapping (FunMap). (A) Logistic genotypic curves for three genotypes at a locus used to simulate growth data with 10 time points under three scenarios, i.e., “Parallel” in which three curves are assumed to be parallel, “Non-parallel” in which three curves neither parallel to each other nor cross over, and “Crossing-over” in which three curves cross over at a certain point. Residual covariance matrix is assumed to follow the SAD(1) structure. (B) – (D) Estimates of genetic effect curves (dotted line), in comparison with true genetic effects curves (solid line), under different simulation scenarios, heritabilities (H2), and sample sizes (n). (E) Empirical power of QTL detection by conventional static mapping for individual time points and FunMap under different simulation scenarios, heritabilities, and sample sizes. The empirical power (given at the upper row) is estimated as the proportion of simulation replicates that are tested to be significant over a total of 1000 simulation replicates, and this “power” becomes false positive rate (FPR) (type I error rate) (given at the lower row) if three simulated genotype curves entirely overlap (i.e., there actually is only one curve specified by a single set of growth parameters). Note that FPR was estimated under different residual variances corresponding to a given heritability level used to calculate the power. The critical threshold at the  = 0.05 significance level for each simulation replicate is determined from 1000 permutation tests. Related to Figures 2 – 6.