BLUPs and Heritability, Mixed Model Tutorial in R

[@jamesjesusjames4810]

Thanks kevin bro, It was really helpful me. As a beginner I was able to calculate blups from augmented design

[@dr.ragill2752]

Thanks for the useful video. I am wondering about the model, I have two-year data of a single location, and the RCBD design was used (replicated three times, and 10 plants were taken from each genotype). So, which model I should use? BLUE or BLUP? And how I should calculate the Heritability? Later, I wanted to do SNP-GWAS. Thanks!

[@georgewanjala4605]

Can I use the same to proceedure in animals without having parents information apart from identity of dams and sires?

[@MrJeroentje666]

Hi Kevin, very helpfull tutorial, thanks a bunch! Question: In your experimental design you have 12 (or 14?) reps (one in every subblock) if I understood you well, so you set your nloc to "14". My data is not balanced, I have more replications for some genotypes and less for others, I am a bit uncertain if I shood set nloc to anything... If I do not use nloc (i.e. set it to 1) my heritability becomes rather low. What would your advice be?

[@maheshbiradar2510]

hello kevin... please share with me the phenotypinc dataframe to know for stepwise analysis

[@endeshawtadesse]

Thank for the video and could you please share me the soft copy of the book you recommended.

[@eustrainroblero729]

can we estimate different type of blup? for example Blup broad , Blup Narrow and Blup intermediate

[@krishnaghimire4139]

The heritability estimates seems to be too high. Can you please conform if the equation you used is the correct one. I have similar root traits (total length, primary root length, lateral root length) and trying to calculate heritability. Are there any other models or equations used for heritability calculation.

[@nikeeshrestha6011]

I think that formula is for narrow sense heritability instead of broad sense as he is not including variance of residuals.

[@wardahmustahsan888]

Hi Kevin This was super helpful! I learned a lot about this. I am new to heritability but this was a great resource to learn. I have a question about your model. Did you ever encounter and error like this: 'Error: number of levels of each grouping factor must be < number of observations' I mean I think I know why because the data I am using is from a p-rep design so there is variation in the reps of the entries so maybe this is effecting the model. There is a temp solution but I don't really like it because all of the variance values become the same. With your data did all of your entries have the same # reps?

[@cindyrouet7822]

Hi Kevin. Thank you for your tutorial, it has been really interesting. I am getting into BLUPs analyses and heritability estimates for cold hardiness in polyploid roses. I am working at the moment with only two environments so I am thinking setting environment as a fixed effect and genotype as a random effect. My response variable is a ratio of two stem measurements taken in the field. I was wondering if you had any experience or suggestion on modelling proportion data with lme4 specifying a beta distribution. Thank you.

[@KevinFalk]

Hi Cindy. Thank you for your kind words. Interesting question. In my experience I have yet to work with a beta distribution however, Sarah is working with nematode counts with a similar distribution issue. I will look deeper into this.

[@cindyrouet7822]

@@KevinFalk Thanks a lot!

[@andrisuheri2503]

Hi Kevin.. have you any suggest others book have already discuss about your topic? And thank for your attention

[@KevinFalk]

I recommend Breeding for Quantitative Traits in Plants by Rex Bernardo.

[@andrisuheri2503]

@@KevinFalk Dear Mr. Kevin.. if you mind could you share a date set as an example..? maybe this very helpfull to me to learning step by step.. many thank & God bless you

[@HunterDriguez]

I appreciate the very helpful video! I was wondering, what is "Entry" in df? Is it a unique plant identifier or do all 12 reps within each genotype share the same Entry number? Maybe I'm completely off.

[@KevinFalk]

Entry is the genotype. It will appear in each replicate.

[@HunterDriguez]

@@KevinFalk Thank you!

[@jizhogjie1205]

Hi, Kevin, I am curious your experimental design? Could you explain a little bit of your design, which concerns the linear model.

[@KevinFalk]

My experimental design was essentially a RCBD using 12 growth chambers (which I call sub-blocks in the video but are essentially reps) and 292 genotypes (of which 12 plants were grown of each).

[@jizhogjie1205]

​@@KevinFalk Another question, did do the ANOVA test to check what's factors(like, Subplot) are significant or not then put then in BLUP ?

[@KevinFalk]

@@jizhogjie1205 yes I did do that

[@buithephong]

Thank you for your nice sharing. Sorry, I tried to get to your github, but could not find the data file (.csv), Can you share your data file? Thank you.

[@KevinFalk]

I cannot share this data file. However, there is a sample data file in my Github located here: /mighster/ARIA2.0/blob/master/05-07-2019_KGF_RawData_iRootonly.csv

[@buithephong]

@@KevinFalk Thank you for your prompt reply. I'm also doing phD in root system traits in soybean (GWAS) and this is very helpful.

[@maheshbiradar2510]

hello you got data file CSV

[@maheshbiradar2510]

if u means please share with me

[@mahadevaiahc7758]

Sir Please provide the methodology for estimation of Pedigree Matrix from Coefficient of Coancestry regards mahadev

[@rizkianjalpujinugroho8244]

hi Kevin, my name rizki and I'm oil palm breeder, can I see your data frame ? thank you very much,

[@KevinFalk]

Hi Rizki. My code and data files are on my Github page, link in the description.

[@TheEbc123]

thanks!

[@georgewanjala4605]

Interesting