Question: PredictABLE R package - HELP building a weight genetic risk score
gravatar for Juliana
5.6 years ago by
Juliana40 wrote:

Hi Dear All,

Please, I need some help. I´m new in Bioinformatics and I really appreciate if someone could just show me some "light".

I´ve been building a weighted Genetic Risk Score (wGRS) in order to evaluate it effect on the prediction of  risk of developing a disease according to the clinical data from my studied population. I have been using RStudio and the PredictABLE R package. 

I was able to build the unweighted Genetic Risk Score, which evaluate a simple risk allele count assuming that all alleles have the same effect ( the genetic variables are coded as the number of risk alleles 0, 1 or 2). But actually I want a weighted risk score that include a sum of the number of those risk alleles multiplied by their beta coefficients (ln from previously OR obtained in GWA studies), and then divided by the total numbers of SNPs included in the model. I am able to do this in excel. But not with PredictABLE.

I should be able to calculate the RiskScore using the following function:

riskScore <- riskScore(weights=riskmodel, data=ExampleData, 
cGenPreds=cGenPred, Type="unweighted")

However I do not understand what the tutorial package means when ask me :

"This vector of beta coefficients should be a named vector containing the same names as mentioned in genetic variants".

They tell me that weights should be the riskmodel (I calculated it using a fitLogRegModel function. But for this I´m obliged to use only the risk allele coding...Or should I consider already the sum of these risk alleles? And about GenPreds that ask me about the genetic variables? What does it mean? I am really lost here.)

Hope I was clear...

Thank you so much...


Arguments from the function:

The vector that includes the weights given to the genetic variants. See details for more informations.


Data frame or matrix that includes the outcome and predictors variables.


Column numbers of the genetic variables on the basis of which the risk score is computed.


Specification of the type of risk scores that will be computed. Type can be weighted (Type="weighted") or unweighted (Type="unweighted").


ADD COMMENTlink written 5.6 years ago by Juliana40

Did you ever get a solution for this? I'm trying to create the vector and I'm also confused what they mean by "named vector containing the same names as mentioned in genetic variants".

ADD REPLYlink written 5.2 years ago by jennavive0020
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1196 users visited in the last hour