Assessing the utility of genomic deep learning models for disease-relevant variant effect prediction