Principal component analysis in R | PCA for genetic diversity assessment using varimax rotation |

This video clearly explains the procedure involved in principal component analysis especially when we are using pca for genetic diversity assessment in plant breeding. Here you will understand why we need to do pca, how pca reduces dimensions, how to use rotation and how to interpret the graphs and results ?. In the core of this video you will find the syntax or the code which you need to run pca in R along with explanation for each and every argument explaining its importance with that i do show a lot of customisation options you need to follow in R to get beautiful, aesthetically appealing plots especially rotated component plot.

Reference Thesis

For those who are interested in theory check out the articles below
~ PCA in R analyticsvidhya
~ PCA step by step
~ PCA one stop shop
~ Learn PCA in 3d

In order to know about rotated component matrix please check the below article
~ Rotated component cross validated

In order to know about the different functions of R
~ FactoMineR
~ prcomp vs princomp
~Add info

00:00 Intro
05:46 Data structure in excel sheet
06:31 Beginner tips
07:11 Importing data
08:42 Scaling
10:36 Adjusting options
11:46 Visualisation packages
13:25 PCA-princomp
14:52 PCA-prcomp
20:46 3d plots
22:55 PCA-FactoMineR
34:04 Judging number of components
38:07 Elbow method
39:21 Rotated components

Script
Dataset

[Guruprasad_A]

Check the script with dataset in the video

[joankibet]

wow wow wow, so helpful, found it when I was really struggling to run PCA analysis. Thank you

[priyankaparihar]

you are amazing.
Thanku for sharing this information in such easy and detailed manner

[ramanathanthusyanthini]

I do not know how to express my thankfulness. I had been worrying about PCA. Fortunately, i found your video and started to watch it. Now i am completely clear. Could please do more video . These are really helpful. I highly recommend your teaching to others. Thank god.

[JyotiKumari-fujs]

Well explained sir .. and in just simple way thankuh so much

[nikamarolt]

Thank you very much for your explanation.

[obadiahrai]

Big thanks to you. This video is a great help to me. 👍👍👍👍

[emilyford]

Thank you so much, it's very useful for a beginner.

[Scholar_shubham]

God type explanation, thank so much sir❤

[deeptitiwari]

Thank you for the informative video. Please make videos on STRUCTURE and GWAS analysis also. The knowledge of statistics is required in all the fields, please try to spread it for everyone.

[poliyugandhar]

Your videos are so simple and informative. Please do the video on QTL mapping. That will be a great

[deepikachandrasekaran]

Sir I am watching your videos regularly..they are very simple and easy to learn... please do video on variability and path analysis using unreplicated data.

[bhumikasinghlodhi]

Commendable work.
Just a doubt, why you have taken nfactor as 7, rpca <- principal(scd, nfactors = 7, rotate = "varimax", scores = TRUE )

[akhildonga]

Simply explain the PCA analysis. but have one query about if we have a number of observations (Characters) that are more (2x) than a number of treatments (genotypes) then what changes do?

[adityapratapsingh]

Sir, the same codes used by you in the video would work well when some variables are categorical along with quantitative variables?

[ikrambashir]

Hello thanks for the very informative video. I have a question. I have data with 2 factors, Genotype and environment (control and stress) and I have almost 40 variables. So, I want to apply PCA. Should I apply PCA separately on control and stress data or should I apply it together? what you suggest???

[abigaila.jubilee]

Thank you very much for your video sir. I'm already converted the data using log 10 (x+1). Can I still use the options (scipen=100)? Thanks

[drbantikumar]

when running below script
>rpca = principal(scd, nfactors = 18, rotate = "varimax", scores = T)


shows following error
Error in principal(scd, nfactors = 18, rotate = "varimax", scores = T) :
could not find function "principal"

please resolve

[ikrambashir]

In Rotated component analysis I am getting this warning. is it normal or should I need to do something
> rpca <- principal(scd, nfactors = 7, rotate = "varimax", scores = T)
Warning messages:
1: In cor.smooth(r) : Matrix was not positive definite, smoothing was done
2: In principal(scd, nfactors = 7, rotate = "varimax", scores = T) :
The matrix is not positive semi-definite, scores found from Structure loadings

[golumworks]

If I just use addEllipses =TRUE, what determines the size of those ellipses? Also, if I specify ellipse.type = “confidence”, what confidence level is used to generate the ellipses? I used factoextra.