Returns a list of example datasets.
Value
A list with the following elements:
QTL
: a tibble of QTL regions, with columnsid
,chromosome
,start
,end
,score
andname
.GWAS
: a tibble of GWAS results, with columnsid
,chromosome
,position
andscore
.DE
: a tibble of differential expression results, with columnsgene
,chromosome
,padj
,log2FoldChange
,start
,end
andlabel
.CAN
: a tibble of candidate genes, with columnsid
,chromosome
,start
,end
,name
andgene_name
.
Details
The dataset used in this example is presented in: Angelin-Bonnet et al., BMC Plant Biology (2023). In this
study, tetraploid potato plants from a half-sibling breeding population were
used to assess the genetic components of tuber bruising. Capture sequencing
was used to obtain genomic information about the individuals, and a
genome-wide association study (GWAS) was performed on 72,847 genomic
biallelic variants obtained from 158 plants for which a bruising score was
measured. The GWAS analysis was carried with the GWASpoly
package. In
addition, expression data was obtained for 25,163 transcribed genes, and a
differential expression (DE) analysis was carried out between 41 low- and 33
high-bruising samples. Finally, a literature search yielded a list of 42
candidate genes identified in previous studies as involved in potato tuber
bruising mechanisms. A subset of the GWAS and DE results, as well as the list
of candidate genes from the literature, are made available in this function.
From the complete GWAS results table, half of the genomic variants with a
GWAS score < 3.5 were randomly selected and consequently discarded, yielding
a dataset with GWAS scores for 35,481 variants. Similarly, half of the
transcribed genes in the DE results table with an adjusted p-value > 0.05
were randomly selected and discarded, yielding a dataset with DE results for
10,671 transcribed genes. This filtering was performed to reduce the size of
the datasets (in accordance with CRAN policies), but ensures that all
significant markers and genes are retained in the datasets. Finally, some of
the candidate genes located on chromosome 3 were removed from the example
dataset for better clarity in the resulting HIDECAN plot, leaving 32
candidate genes. The QTL mapping data was randomly generated for illustration
purposes, as no QTL mapping was performed in the original study.