What is the most efficient way to identify candidate genes and their functions for GWAS results in cannabis when the current reference genome only shows LOC IDs and incomplete functional annotations? Can alternative fully annotated genome files from other species be considered for this purpose?
Thanks in advance
The most efficient way to identify candidate genes and their functions for GWAS results in cannabis with an incomplete reference genome is to employ a combination of strategies:
Use functional annotation tools: Despite having an incomplete reference genome, you can still use tools like Blast2GO, InterProScan, and EggNOG-mapper to infer gene function based on sequence similarity, conserved domains, and gene ontology (GO) terms.
Cross-species comparisons: You can use closely related species with well-annotated genomes, such as Arabidopsis or hop, as a reference for functional annotation. Compare the cannabis genes with their homologs in these species using BLAST or similar sequence alignment tools to identify conserved functions.
RNA-Seq data: Generate RNA-Seq data from different tissues and conditions to identify differentially expressed genes (DEGs) associated with your GWAS results. This can help you prioritize candidate genes based on their expression profiles and may provide insights into their functions.
Co-expression networks: Use gene co-expression network analysis to identify groups of genes that are co-expressed with your candidate genes. This can help you infer their functions based on the known functions of other genes in the network.
Experimental validation: Perform functional validation experiments, such as gene knockouts, knockdowns, or overexpression, in a suitable model system to confirm the role of candidate genes in the trait of interest.
While alternative fully annotated genome files from other species can provide useful information for inferring gene function, it is important to consider the evolutionary distance between the species. Comparing cannabis genes with those of closely related species will yield more accurate and meaningful results.
Qi's answer is a great summary of approaches to post-GWAS functional studies when you know the gene or genes likely mediating the GWAS signals' effects. But if the cannabis genomic architecture is anything like human, I suspect most GWAS significant variants lie in non-coding regions, meaning the first challenge is identifying the genes likely to be mediating the effects of these genetic signals. This is a big challenge that continues to stymie post-GWAS functional studies in humans even though many epigenomic, chromatin conformation, and eQTL resources have been well developed across most tissues.
There have been a lot of approaches developed to nominate such candidate genes for GWAS signals, and some of them may be adaptable for your situation where annotations are poorly developed. Take a look at Figure 2 of this paper for examples of such approaches: https://www.nature.com/articles/s41588-023-01314-0
Good luck!
Thank you, Qi and Jason, for the great information . They were truly helpful. Much appreciated!