I have a population of a dioecious species with significant phenotypic variation and want to select individuals from this population. What statistical methods can I use to perform multi-trait selection or dissimilarity studies with data of individual plants? If have a R package is better yet.
Dear Hugo Gabriel Peres Your question is highly interesting. To my opinion, before performing multi-trait selection you will have to pay attention on whether the traits in question are uncorrelated or correlated, and whether your population consists of equal number of male and female individuals. If traits are genetically uncorrelated, for practicing 5% selection intensity you will have to select 37% best individuals for each trait (if you are concentrating on 3 genetically uncorrelated traits). If your population consists of equal number of male and female individuals, it's well and good. However, if there is significantly unequal number of male and female individuals, then your effective population size will drastically come down! Keeping these facts into consideration, you may proceed further as suggested by Maurice Ekpenyong
Thanks!
If your trait is defined, then compare the dioecious plants using your identified trait as parameter. Multiple t-test can give you good results and selection using 95% confidence interval or adjusted p-value otherwise called the q value is very useful too. I perform these tests with GraphPad Prism.
You could look at the "trova" function implemented in R, which allows intraspecific trait variability. It is described in :
De Bello et al. (2013) Which trait dissimilarity for functional diversity: trait means or trait overlap? Journal of Vegetation Science.
Hope that helps.
GA2
Dear Hugo Gabriel Peres Your question is highly interesting. To my opinion, before performing multi-trait selection you will have to pay attention on whether the traits in question are uncorrelated or correlated, and whether your population consists of equal number of male and female individuals. If traits are genetically uncorrelated, for practicing 5% selection intensity you will have to select 37% best individuals for each trait (if you are concentrating on 3 genetically uncorrelated traits). If your population consists of equal number of male and female individuals, it's well and good. However, if there is significantly unequal number of male and female individuals, then your effective population size will drastically come down! Keeping these facts into consideration, you may proceed further as suggested by Maurice Ekpenyong
Thanks!
Hi,
I agree with Maurice Ekpenyong, you have to follow his advice.
Regards
hi
i think you can use t.test or other thest and if you want to chek the relationships between each of them use regresions method or pearson cortusis .
Article Phenotypic characterization of recurrent selection S1 Papaya...
I think this paper solved the question
- Badges
- Science topic
- Similar topics
- Agricultural Science
- Horticulture