I have un-replicated mean data of various grain yield traits from two consecutive years. I want to estimate the broad sense heritability from these data. Can I estimate? Kindly give your kind suggestions. If yes then please share me an example or any link?
you can estimate it.
here you are my study i estimated it
Thesis Genetic Studies on Some Vegetable Crops (Corchorus olitorius...
good luck!
I think you can estimate error variance from years as like as error variance estimated from replicated trials. similarly, you can also get the phenotypic variance and genotypic variance. So, i think you can estimate heritability
Dear Ehab,
I saw your thesis thoroughly. You have used the replicated data for estimation of heritability. By the way, which software you have used for the estimation of heritability?@Ehab M. B. Mahdi
Heritability normally requires an estimate of error variance to separate genotypic variance from phenotypic variance. However, if you have genomic information for individuals (e.g. derived from SNP chips) then you can derive a genomic relationship matrix and estimate genomic breeding values from single rep trials - and you should be able to estimate heritability from the output.
BSH = (Total geneticvariance/phenotypic variance)*100
If the same lines (G) are evaluated in two years (Y) and are truly random and large in number, then you have two sources of variation: G and Y. The GxY is your error (E) SOV and your ANOVA is
SOV df EMS
Y y-1
G g-1 s2e + 2s2g
E (y-1)(g-1) s2e
You can easily obtain s2g and s2phenotype from the ANOVA and then estimate the BSH. Be warned, however that this is a rough estimate and should be cautiously applied or extrapolated in further studies.
Error variance is required to estimate heritability in broad sense. So in case of unreplicated dats, it is not possible to estimate heritability.
Repeating a study in another year is replication in time and, therefore, (g-1)(y-1) is the error term in this case. If the study were conducted in one year with 2 reps. (g-1)(r-1) would be the error term. In principle, there is no difference between the two approaches. Error is simply the variation remaining in the experiment when all the known sources of variation have been deducted. That is the reason it is called "residual"
if you have some genotypes replicated in the same conditions that may be used to calculate environmental effects. moving average analysis could be another option to compare the performance.
Dear Sir,
Thanks for the informative and useful answer M. A. B. Fakorede
Yes you can if some check lines have been replicated. Efficiency of these check-plot designs is discussed in
Article Efficiency of check-plot designs in unreplicated field trials
I have encountered the same problem before. For perennial woody plant germplasm conservation, usually only one plant of each cultivar is preserved. To calculate heritability, we can use the R package 'lme4' to calculate the heritability by duplicating the trait data from different locations or years.
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology
- Research Papers