Hello, I have a large accession number (n=200). I will screen them for blast disease resistance and then conduct the GWAS. I will conduct the experiment in greenhouse where I have limited space. My question is, which experimental design would be suitable for phenotyping when you have limited space and large sample number? I read that augmented design can be an option. Please share some ideas.
Augmented design is the most suitable in your case. add checks(example four) to your 200 that you sould divise in 10 block( each one 20 accession +4 checks). checks sould be replicated in each block.
As Mr. Tidiane said, the Augmented Blocks Design is the most suitable for your situation. But be careful about the checks. You should know their features well, and chose the ones that are most similar to your accessions. Just remembering: the experimental errors will be estimated based on checks' variation, that's why the choice of them is so important.
Bests
- This is a difficult question! Augmented designs are a good option in experiments with large number of entries and few resources and is useful for selecting superior and inferior genotypes but not too good discriminating intermediate ones. On the other hand GWAS needs precise phenotypic information in order to achieve good asociations between genes and phenotypes. Depending on the variables you work with, also affects wich design is more suitable, an easy to measure unnaffected by enviroment variable could give acceptable results in augmented designs, but if this is not the case, you should use more powerful replicated designs, like incomplete block designs (alfa lattice) or a replicated evaluation in groups of individuals instead. In my experience is better to reduce the size of the plot in order to increase the number of plots in the same space and use more powerful designs.
- Hello Jesus, thank you for your answer. Actually I have very little space and I am going to do the experiment on trays which can accommodate 96 different genotypes. I can have a double number of plants of each genotype. I think the one you mentioned about a replicated evaluation in groups of individuals seems feasible in my case. Can you please explain this procedure or do you have any reference for that?
I would suggest you to consider the inheritance of trait under study before designing the GWAS experiment. For example, quantitative trait needs more replication of accessions within an experiments as compared to qualitative traits. Good luck.
I also agree with Jesus. You can do row-plot and even reduce the length plot (for example 2 or 3m). 3 reps will be ok if you have space if not 2 reps can help I think. All the best.
Two dimensional row-column design would be apt for your case. However, make sure to maintain uniform condition throughout the experimental run and measure the values at same time. More accurate the experimental error is the better chance to screen potential lines. Hope this helps.
- Badges
- Science topic
- Similar topics
- Microbiology
- Assays