Computational protocol: Can natural variation in grain P concentrations be exploited in rice breeding to lower fertilizer requirements?

Similar protocols

Protocol publication

[…] The 700K SNP genotyping data [] and the software Tassel 5 were used for the association study. The SNP data were modified as follows: heterozygous SNPs were first set to missing values. They were further filtered using 2% as the minimum allele frequency and 85% as the minimum count. In the end, 261K SNPs remained for the analysis. Principal component (PC) analysis was performed using an R package (R version 3.1.1) called Genome Association and Prediction Integrated Tool (GAPIT) to control for population structure and the obtained PC data was imported to Tassel for GWAS analysis. As the emphasis of the experiment shifted (see ) we attempted to minimize the type-II error (false-negatives) and selected peaks with thresholds of 1.0E-04 (for mixed linear model, MLM) or 1.0E-05 (for general linear model, GLM) for further detailed investigation. Using linkage disequilibrium (LD) analysis in HaploView 4.2, QTL regions were defined by markers being linked to the peak marker with an R2 ≥ 0.65 (within 80–360 Kb region around the peaks). For the haplotype analysis markers in that LD block with low P value in GWAS were included. Accessions with missing yield data or a HI below 0.30 in both replicates were excluded in the genome wide association analysis to avoid potentially confounding effects (insufficient dilution with low grain yields). In total 150 accessions were included in the GWAS using the indica diversity panel. In the case of the broad association panel the 44K SNP dataset [] was used and associations detected using Tassel 5 with 219 accessions of the indica, aus and tropical japonica subpopulations. […]

Pipeline specifications

Software tools GAPIT, Haploview
Application GWAS
Organisms Oryza sativa
Chemicals Phosphorus