3. The left side of the RESULT page provides information about
boosting performance. The prediction performance by araGWAB is expected to be higher than that of GWAS alone or
of a random network for the user-input genes known for the
phenotype (see Note 6). The optimal p-value threshold (vertical line) is used to filter for SNPs during the boosting process
to generate the final candidate genes. araGWAB selects the
optimal threshold where araGWAB reaches maximal prediction
performance. The final list of candidate genes by araGWAB is
based on the optimal threshold.
4. The final list of candidate genes with araGWAB scores (see Note
7) is available for download.
4 Notes
1. Before phenotyping your accessions, we recommend ensuring
the genotypes of the accessions obtained from the stock center.
A previous study examined 5965 accessions and concluded that
286 of them deserved special attention as being potentially
misidentified [26]. The authors classified the accessions into
three categories: green list—well-validated, yellow list—slightly
suspicious, and red list—mislabeled. It is highly recommended
to use only the accessions in the green list.
2. The first generation of the accessions from the stock center
might not be in good shape, so we recommend using the seeds
from the second generation to reduce the risk of confounding
effects during phenotyping. The seeds shipped from the stock
center sometimes neither germinate nor flower. Vernalizing the
seed for three weeks or longer may aid in germination.
3. It is very important to strictly follow the data file formats. The
scripts will only be able to process input files that are in exactly
the correct format; otherwise, they will show error messages.
4. Phenotyping accessions is the most time-consuming and laborious step of GWAS analysis. Thus, it is important to record
phenotypes in a way that can be revisited, such as images of the
plant rather than just the number of leaves or the length of
seedlings. If a significant association between genotype and
measured phenotype is not detected, you might need to examine other related phenotypes as input. Suppose the other input
phenotype is leaf shape, you can qualitatively assess phenotype
accessions by checking whether the leaves are serrated. If you
cannot detect any candidate SNPs associated with serrated
leaves, you will end up with no candidate genes. However, if
you have leaf images as phenotype data, you can revisit images
and re-phenotype them in a more quantitative manner, such as
Genome-Wide Association Studies in Arabidopsis
207
boosting performance. The prediction performance by araGWAB is expected to be higher than that of GWAS alone or
of a random network for the user-input genes known for the
phenotype (see Note 6). The optimal p-value threshold (vertical line) is used to filter for SNPs during the boosting process
to generate the final candidate genes. araGWAB selects the
optimal threshold where araGWAB reaches maximal prediction
performance. The final list of candidate genes by araGWAB is
based on the optimal threshold.
4. The final list of candidate genes with araGWAB scores (see Note
7) is available for download.
4 Notes
1. Before phenotyping your accessions, we recommend ensuring
the genotypes of the accessions obtained from the stock center.
A previous study examined 5965 accessions and concluded that
286 of them deserved special attention as being potentially
misidentified [26]. The authors classified the accessions into
three categories: green list—well-validated, yellow list—slightly
suspicious, and red list—mislabeled. It is highly recommended
to use only the accessions in the green list.
2. The first generation of the accessions from the stock center
might not be in good shape, so we recommend using the seeds
from the second generation to reduce the risk of confounding
effects during phenotyping. The seeds shipped from the stock
center sometimes neither germinate nor flower. Vernalizing the
seed for three weeks or longer may aid in germination.
3. It is very important to strictly follow the data file formats. The
scripts will only be able to process input files that are in exactly
the correct format; otherwise, they will show error messages.
4. Phenotyping accessions is the most time-consuming and laborious step of GWAS analysis. Thus, it is important to record
phenotypes in a way that can be revisited, such as images of the
plant rather than just the number of leaves or the length of
seedlings. If a significant association between genotype and
measured phenotype is not detected, you might need to examine other related phenotypes as input. Suppose the other input
phenotype is leaf shape, you can qualitatively assess phenotype
accessions by checking whether the leaves are serrated. If you
cannot detect any candidate SNPs associated with serrated
leaves, you will end up with no candidate genes. However, if
you have leaf images as phenotype data, you can revisit images
and re-phenotype them in a more quantitative manner, such as
Genome-Wide Association Studies in Arabidopsis
207