measuring the level of serration and the width of the leaves
when the leaves are serrated. This will save an enormous
amount of labor and research time.
5. There is no rule of thumb for multiple hypothesis corrections.
Selecting one method over the other is entirely dependent on
the GWAS result and the nature of the data. You can select a
method of multiple hypothesis correction for the given GWAS
results based on prior knowledge. For example, you could
choose a method that retrieves the largest number of SNPs or
genes already known to be involved in the phenotype.
6. Augmentation of the given GWAS signals using araGWAB may
not be effective for various reasons. A factor that likely influences the effectiveness of the network-based GWAS boosting is
the functional gene network. The version of araGWAB as of
April 2019 uses AraNet version 2 [27] as a base network.
Although AraNet v2 is one of the most accurate and comprehensive networks of Arabidopsis genes, it still has uneven sensitivity and specificity across different biological processes. If an
underlying biological process for the GWAS phenotype is not
well modeled in the base network, the efficiency of networkbased augmentation of GWAS signals will be limited. You can
evaluate the network prediction power for each biological process with a companion web application of AraNet v2 (https://
www.inetbio.org/aranet/), the usages of which are described
in a related protocol [28].
7. The final score (araGWAB score) for each candidate gene is
simply the sum of edge weights, which may not be interpreted
as a standard metric of significance or confidence level. Therefore, we recommend using the araGWAB score as a relative
value for prioritizing candidate genes.
Acknowledgments
This work was supported by a National Research Foundation of
Korea (NRF) grant funded by the Korean Government (MSIT)
(NRF-2018M3C9A5064709, NRF-2018R1A5A2025079) to I.L.
We thank Sang-Dong Yoo and Geundon Kim for discussions and
sharing Arabidopsis phenotype data.
References
1. Lipka AE, Kandianis CB, Hudson ME, Yu J,
Drnevich J, Bradbury PJ, Gore MA (2015)
From association to prediction: statistical
methods for the dissection and selection of
complex traits in plants. Curr Opin Plant Biol
24:110–118. https://doi.org/10.1016/j.pbi.
2015.02.010
2. Brachi B, Morris GP, Borevitz JO (2011)
Genome-wide association studies in plants:
the missing heritability is in the field. Genome
Biol 12(10):232. https://doi.org/10.1186/
gb-2011-12-10-232
3. Korte A, Farlow A (2013) The advantages and
limitations of trait analysis with GWAS: a
208
Tak Lee and Insuk Lee
when the leaves are serrated. This will save an enormous
amount of labor and research time.
5. There is no rule of thumb for multiple hypothesis corrections.
Selecting one method over the other is entirely dependent on
the GWAS result and the nature of the data. You can select a
method of multiple hypothesis correction for the given GWAS
results based on prior knowledge. For example, you could
choose a method that retrieves the largest number of SNPs or
genes already known to be involved in the phenotype.
6. Augmentation of the given GWAS signals using araGWAB may
not be effective for various reasons. A factor that likely influences the effectiveness of the network-based GWAS boosting is
the functional gene network. The version of araGWAB as of
April 2019 uses AraNet version 2 [27] as a base network.
Although AraNet v2 is one of the most accurate and comprehensive networks of Arabidopsis genes, it still has uneven sensitivity and specificity across different biological processes. If an
underlying biological process for the GWAS phenotype is not
well modeled in the base network, the efficiency of networkbased augmentation of GWAS signals will be limited. You can
evaluate the network prediction power for each biological process with a companion web application of AraNet v2 (https://
www.inetbio.org/aranet/), the usages of which are described
in a related protocol [28].
7. The final score (araGWAB score) for each candidate gene is
simply the sum of edge weights, which may not be interpreted
as a standard metric of significance or confidence level. Therefore, we recommend using the araGWAB score as a relative
value for prioritizing candidate genes.
Acknowledgments
This work was supported by a National Research Foundation of
Korea (NRF) grant funded by the Korean Government (MSIT)
(NRF-2018M3C9A5064709, NRF-2018R1A5A2025079) to I.L.
We thank Sang-Dong Yoo and Geundon Kim for discussions and
sharing Arabidopsis phenotype data.
References
1. Lipka AE, Kandianis CB, Hudson ME, Yu J,
Drnevich J, Bradbury PJ, Gore MA (2015)
From association to prediction: statistical
methods for the dissection and selection of
complex traits in plants. Curr Opin Plant Biol
24:110–118. https://doi.org/10.1016/j.pbi.
2015.02.010
2. Brachi B, Morris GP, Borevitz JO (2011)
Genome-wide association studies in plants:
the missing heritability is in the field. Genome
Biol 12(10):232. https://doi.org/10.1186/
gb-2011-12-10-232
3. Korte A, Farlow A (2013) The advantages and
limitations of trait analysis with GWAS: a
208
Tak Lee and Insuk Lee