I have more than 8000 SNPs identified from GBS approach of 90 diverse cultivars. To train myself for association mapping study, I would like to start with disease phenotypic data to apply for GAPIT. The question here I would like to ask and need help of experienced experts or who faced same issue is , how can I prepare my raw phenotype data of disease resistance? For example I have 20 resistant and 70 susceptible cultivars, do I need to code as it is ‘r’ for resistant and ‘s’ for susceptible lines, or i should code it in numerical only? Please kindly provide your advice and suggestions, I would very appreciate your help
I use comments. Thanks
With the best wishes
Mr Rezaei Ahvanooei.
Sovetgul Asekova The problem is that you hardly can put diverse reaction of 90 cultivars to the pathogen into simple groups R or S.
Some of them can be resistant since seedlings, some - in adult stage only, some - have HR response, others - can show reduced disease development, etc. Plant reaction can be pathogen strain-specific, temperature, pathogen load, plant stage - depended. It is better to use several traits related to resistance as independent data for analysis.
Best wishes
Alex
Instead of 0,1 for disease resistance/susceptibility, use a scale (say 1 to 5 or 1 to 10) based on the disease. GAPIT cannot handle 0/1 data since it has to use logistic regression. AM in plants is mostly on continuous traits.
Do you have reps? if so makes things easier by using an average. Look at few of my papers if you need a better understanding.
A cautionary note: there are some more basic issues you need to consider here. Typical plant genomes (and when you describe "cultivars" I presume you are working with a plant species) contain 25,000-plus genes, so with 8000 SNPs from your GBS you are only capturing a small fraction of the potentially-causative genes in the first place. Moreover, it takes quite a few SNPs per gene, including regulatory regions, to achieve a high likelihood of capturing association with the actual causative SNPs. This is due both to recombination and the different coalescent histories tracked by the origins of different SNPs. Thus, you are orders of magnitude short on the number of SNPs you would need to detect genome-wide associations.
One potentially valuable use of your data would be to estimate the amount of population structuring across your set of cultivars. The amount of population genetic structure will affect your ability to distinguish causative vs. nonfunctional SNP-trait associations with an appropriately-sized set of SNPs, and the types of control you would need to apply to factor out effects of kinship and population structure.
I sincerely THANK YOU ALL for valuable comments, these are indeed useful to prepare my Raw data, and i am getting to start
Best regards,
Sovetgul
Sovetgul, i agree with Alex response.
Some time we do blunders in scales( 0/1). but still many people believe in this criteria.
For me, Sujan and Alex comments are important. Plant mostly shows a disease resistance scale which could be converted into disease index, AUDPC etc. and could be used for AM. I doubt if you are really doing it in right way. Thanks
- Badges
- Science topic
- Similar topics
- Agricultural Science
- Agronomy
- Crop Science
- Crop