Principle component analysis
0
1
Entering edit mode
10 months ago

while performing PCA to reduce the dimension of the variable using "Factomine R" package in R. I want to extract the each individual gene contribution towards principle component. using this function "var$contrib:" I extracted the file but I unable to set the cut off I mean upto which cutoff I have to select the genes.

I am also attaching the code:

library("FactoMineR")
library("factoextra")


data2 <- read.csv('t_data_processed_1_scale1.csv', header = TRUE)
res.pca <- PCA(data2[,1:X])
print(res.pca)
var <- get_pca_var(res.pca)
var
eig.val <- get_eigenvalue(res.pca)
eig.val
write.csv(eig.val,"feature_eigval_tissue.csv")

fviz_contrib(res.pca, choice = "var", axes = 1, top = 10)

featureCos2=var$cos2
write.csv(featureCos2,"featureCos2_tissue_new.csv")
featureContrib=var$contrib
write.csv(featureContrib,"featureContrib_tissue_new.csv")
PCA cut-off • 416 views
ADD COMMENT
0
Entering edit mode

you can slice featureCos2 or featureContrib as a regular dataframe. Example:
df[df$X>0.8,]
In the example above, you are selecting the rows of df that have a value greater than 0.8 in the column X.
Also try to format your question, using the code typography. Try to have each line of code in a different line, it makes it easier for the rest people to read and understand. Cheers!

ADD REPLY
0
Entering edit mode

Thank you very much for your reply. df$X>0.8 It is mandatory or we can set the cut off 0.7 or 0.6 as well?

ADD REPLY
0
Entering edit mode

I have two datasets having having higher featurecontrib value in one dataset is .97 while in another dataset .47. how to set the cut off for each dataset.

ADD REPLY
0
Entering edit mode

You can set the cutoff you like, 0.8 was just a random number I used for the example. The cutoff you are applying is arbitrary so you decide which is the threshold you want to use.

ADD REPLY

Login before adding your answer.

Traffic: 1592 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6