Multivariate D-Square statistics in R/Tocher method of clustering for genetic diversity assessment
HTML-код
- Опубликовано: 12 июл 2024
- Hi there in this video, I have shared the complete procedure involved in Multivariate D-square statistics in plant breeding, using Mahalanobis squared euclidian distance for assessment of genetic diversity in plants.
This is a second order statistics not the most accurate one as compared to Metroglyph analysis but can be used for more number of individuals, with replications only
~ Here I am using the biotools package developed by Prof. Anderson da Silva to do clustering based on Tocher and Modified Tocher Method along with the inter and intra cluster distances.
~ The dataset used here is from
sites.google.com/site/tnausta...
~ Cross verify the results if you think about credibility of the analysis
drive.google.com/file/d/1DcuE...
Both of them are from TNAU stat
0:00 Introduction
3:18 Coding
Script
docs.google.com/document/d/1-...
Check the script in the description.....
Thank you for the video , it was really informative and simple to execute 😇😇
very good hard work and wonderful video thanks bro
sir, what is the script for find out the cluster mean value from tocher method
Just a doubt..Can we do univariate analysis for clustering when we have n number of genotypes studied for say m number of traits?
Univariate = one variable,
Multivariate = Multiple or more than one variable,
In our context,
Variable = Trait.
Thank you for this amazing video. I have one doubt that how to find cluster mean table using R ?
If the output was in list. we can use the aggregate function (refer last part of my cluster analysis video), unfortunately it's not in list. so it's difficult to use aggregate function because we need to create list manually and it takes more time, so add an extra columns before genotypes and fill it with cluster membership (1,2,3. .......) Whatsoever for you and sort them based on cluster membership Column and calculate mean by adding a extra row at the end of each cluster.
is it possible to analyze over location lattice design experiment using the explained script?
If you have replication yes
Sir I have found error in code manova calculating part, the r studio showed error in is.factor(x) 'Genotype' not found
Is there any missing genotypes ?
Please do check the name of genotype column in your dataset ?
Hi, everyone don't hesitate to ask your doubts here...... I will try to answer them.
Sir Please tell me how to make image of inter cluster and intra cluster
Hello sir, I found error in summary.manova(mod) : residuals have rank 17 < 18
I have 18 characters. Please suggest what does that mean? And what I have to do ?
stackoverflow.com/questions/39412865/error-in-summary-manova-residuals-have-rank-order-deficiency
Read the above article.
The video was very clear. I have one error while executing the code summary(mod), and it was "Error in summary.manova(mod) : residuals have rank 3 < 8", can you please help me in rectifying this error.
stackoverflow.com/questions/39412865/error-in-summary-manova-residuals-have-rank-order-deficiency
Can we get p- value of each cluster based on mahalanobis distance or can we get p-value of each cluster in LDA analysis?? If yes put the code
No
how can we do D2 of two environment data (as pooled)/ for example control and drought environment, i want to calculate d2 clustering for pooled data involving both the envs. Kindly tell me.
Check this out once in STAR, PB tools.
There we may get an option.
It's better to do it separately for both environment.
Dear The Outlier, Thank you for the very good video. I am facing a problem. The total number of clusters formed using this code is only 2. While the same data was analysed using windowstat the number of clusters are 25. why this difference?
In both the methods also you are getting 2 clusters.
@@Guruprasad_A yes both tocher and modified tocher
Can you able to check distance matrix sir, whether both are some or not.
@@Guruprasad_A hii
Sir, I'm getting error in D2.dist step. I did as said in video but I'm getting error saying dimension is incompatible, I have 12 characters and 18 genotypes, please help...
I think there is a problem with number of variables, which you have considered for calculating variance and covariance matrix and the number of variables in average data.
Thank you
Sir can you suggest how i can make cluster diagram using any other software if not possible in R. Please also tell how i can make tocher based dendrogram in R
You can use window stat software if you have it
Sir, in the R software while doing linear discriminant analysis, it is showing that figure margin is too large. Is there any solution?
Try taking out one variable at a time and re do analysis until u find out the variable which might be that error…
@@Guruprasad_A Thank you sir, will try and let u know....
@@Guruprasad_A Sir I tried the method u said, but the same error is happening, it is showing plot.new (): figure margins too large
Can we define the number of clusters in tocher's method? I am getting 3 clusters but if i want 4 clusters, how would i do that?
You can opt for k means cluster
Or try alternative tocher method
Thanks author to deliver a valuable video, could you tell me how to create Fig.3 (Inter cluster and intra cluster distance...) in the end of the clip? Thanks again
Imagine you have got 5 no of clusters then draw Pentagon which will be having 5 sides at intersection you have to draw a circle so totally 5 circles, later label the 5 circles serially within them mention the intra cluster distance. Later connect all the circles then for example on the line which connects the 5th and 3rd circle (which represents inter cluster distance between 3rd and 5th cluster) mention inter cluster distance between 3rd and 5 th cluster in the same way do this inbetween other circles.
@@Guruprasad_A Thanks for your prompt reply. Did you mean we need to create manually?
Yup
@@truongphu7407 How to create manually?
Hello sir code workig till covar calculation but when i run imp
It means there is a lot of correlation in the variables of your dataset. Please check the correlation among the variables and try to remove those variables..
For example you find correlation in days to first flower and days to fifty percent flowering etc...
Bro how to plot the dendrogram we are getting the clusters na how to plot them can we get that process
Dendrogam is not possible with tocher method.
Thank you so much sir for this video. I would like to ask you about the diagram of tocher method?? How can we make that?
We can't make it, it won't be that much appropriate because it's not a hierarchical method of clustering
@@Guruprasad_A how you make last in video. diagram of 169 genotype by tocher method
Any R code for plotting cluster diagrams with the Tocher method?
No, sorry
If any reference code is there for the same...plz update here
@@sukumartaria3582 ok
How to draw dendogram after modified tocher method ..plse help
Not possible in this package.
Isn't it necessary to scale the data?
I think we should scale both dv matrix and averaged data frame
Calculating paired Mahalonobis distance is kind of scaling only.
@@Guruprasad_A
Will it be valid if i use:
mod
As per the information what I got from the original author, I did made this video. I hope you follow what I have shown in video
I hope this helps you to understand scanning is as good as calculating Mahalonobis distance stats.stackexchange.com/questions/210155/mahalanobis-distance-and-feature-scaling
towardsdatascience.com/mahalonobis-distance-and-outlier-detection-in-r-cb9c37576d7d
Error in model.frame.default(formula = dv ~ as.factor(genotype) + as.factor(Repl), :
variable lengths differ (found for 'as.factor(genotype)')
In addition: Warning message:
In xtfrm.data.frame(x) : cannot xtfrm data frames
>
i am facing this error
place help to solve this problem
That error is because, the number of genotypes and replications are not same.
Check some of them might be missing
@@Guruprasad_A sir i am beginner in the Programming can you tell me how to read the error
@@yellankinaveen2496 check if all the genotypes and replications are present are not
How to generate dendrogram using tochers method in r studio sir
Cluster diagram is not possible in this package
Sir, can you please share how to do the analysis without replicated data ( augmented design)
We can't do tocher method of clustering for non replicated data.
Try Ward min variance method insted, as shown in my cluster analysis video
Thank you sir 👍
@@Guruprasad_A why Ward minimum variance method
@@RamendraSarma It's subjective you can use any other methods such as average, complete etc... But most often used method is ward minimum variance method when it comes to hierarchical clustering in genetic diversity studies using morphological data.
Kindly suggest when no. Of genotypes are less than no. Of traits, these codes are not giving results. It gave best results when genotypes were either more in number or equal to no. of observations.
Not soo, i think.
@@Guruprasad_A can we work for less no of genotypes and more observed traits?? If so then I have to check my data
@@asingh9317 Sure, but inorder to study we need more genotypes.
@@Guruprasad_A got it...thanks alot🍁🍁
Sir, help me to solve this error
Error in is.factor(x) : object 'Genotypes' not found
Check the name of genotypes column in Excel sheet and confirm that same as mentioned in R studio also.
Remember R is case sensitive
Both are same sir even though it is showing the same error
how to f ix this problem please i got same probllem
Check the name of the genotypes column what is there in Excel and in our video they have to be same.
How did you solve that sir.. Even I have the same error though everything is perfect...
How to analyse D square with 2 year of data of fied trail sir
Take multiple year data as replication.
Sir can you explain elaborately, how to do cluster means
Group them individually / seperate the groups in different Excel sheet then take average.
Hello sir...
Sir i didnt get how to find cluster means...can u explain me sir??
Need to calculate manually, by seperating clusters in an Excel sheet.
Sir, Separating clusters means: of single trait the genotypes present in the one cluster have to take the average ???
Add one more column to data mention the cluster number, then sort according to cluster number and take average.
Ok sir...got it...Thank you sir...
please mention code for last cluster diagram
We can't get a cluster diagram in this package. But for the last diagram which depicts inter and intra cluster distance we need to create manually in
Microsoft power point.
thanks for the video.
I am getting error after executing manova command,
Error in `[[
I think it's a problem with your dataset or you might have mis selected the variables.
@@Guruprasad_A I have check the data many times but couldn't get the mistake. can you help me in finding the mistake
@@bhumikasinghlodhi6632Send the dataset, I am quite busy as of now, but I will try to clarify it by next Wednesday.
Please provide me your email
gthings1597@gmail.com
Sir how we draw the diagram of tochers clustering diagram please help me
In this package we can't generate a cluster diagram.
@@Guruprasad_A is there any package sir
@@dibsohbordoloi7952 Nope, as per my knowledge
sir, can you make a video on metroglyph analysis
As of now I am busy with my course work, I look into it later.
I used to visualise clustering using cytoscape......just try this.....
Ok
Hi, Is there any introductory tutorial on the same. Like how it is done in Cytoscape.
@@SonuLangaya well just installs Cytoscape and transforms the matrix according to individual interaction. directly fetch the file and you will find the tocher graph according to inter and intra values
Use “~” tilde not the “-” after dv~
Cluster mean is where? Bro
You have to calculate manually by grouping the varieties or treatments separately based on their clusters identity in an excel sheet.
Thank you Sir for this very important video. I have got a problem while I was doing my data on percentage contribution. The error said:
Error in solve.default(cov) : system is computationally singular: reciprocal condition number = 1.54415e-19
How can I amend/correct it? Please, give me your valuable comment.
It's seems your data is having too much correlation between the variables, please reduce it and see.
@@Guruprasad_A Thank you very much!
@@Guruprasad_A how to rectify that?
can you share the codes?
Sure give me your email I D
or Just send me a message to gthings1597@gmail.com
Where is as.factor function ?
It's built in base r package.
Please can you send codes
Check the description