Warning message clarification

Question

I'm using SNPassoc R package for finding association between data SNPs and a continuous variable outcome. I run the analysis and I got the results; however, I got warning message which is:

Warning in terms.formula(formula, data = data) :
  'varlist' has changed (from nvar=3) to new 4 after EncodeVars() -- should no longer happen!

my model is:

model <- WGassociation (continuous variable ~ covariate +covariate+ covariate  ,data= data)
model

I don't know what it means and should I worry about it or ignore it? Can you please help me?

Answer 1

This warning message is coming glm, which is used by SNPassoc::WGassociation, see this line on GitHub.

Warning message is saying that it is dropping some variable as it is a linear combination of other existing variables in the model.

To reproduce this warning try below example:

# data
x <- mtcars[, 1:4]

# run model, all good
glm(mpg ~ ., data = x)
# Call:  glm(formula = mpg ~ ., data = x)
# 
# Coefficients:
#   (Intercept)          cyl         disp           hp  
# 34.18492     -1.22742     -0.01884     -0.01468  
# 
# Degrees of Freedom: 31 Total (i.e. Null);  28 Residual
# Null Deviance:        1126 
# Residual Deviance: 261.4  AIC: 168

Now add useless combo variable which is constructed from existing variables.

# make a combo var
cyldisp <- x$cyl + x$disp

# run model with combo var, now we get the warning
glm(mpg ~ . + cylmpg, data = x)
# Call:  glm(formula = mpg ~ . + cyldisp, data = x)
# 
# Coefficients:
#   (Intercept)          cyl         disp           hp      cyldisp  
# 34.18492     -1.22742     -0.01884     -0.01468           NA  
# 
# Degrees of Freedom: 31 Total (i.e. Null);  28 Residual
# Null Deviance:        1126 
# Residual Deviance: 261.4  AIC: 168
# Warning message:
#   In terms.formula(formula, data = data) :
#   'varlist' has changed (from nvar=4) to new 5 after EncodeVars() -- should no longer happen!