CLU, PICALM and CR1 were identified as genetic risk factors for late onset Alzheimer's disease (AD) in two large genome wide association studies (GWAS) published in 2009, but the variants that convey this alteration in disease risk, and how the genes relate to AD pathology is yet to be discovered. A next generation sequencing (NGS) project was conducted targeting CLU, CR1 and PICALM, in 96 AD samples (8 pools of 12), in an attempt to discover rare variants within these AD associated genes. Inclusion of repetitive regions in the design of the SureSelect capture lead to significant issues in alignment of the data, leading to poor specificity and a lower than expected depth of coverage. A strong positive correlation (0.964, p<0.001) was seen between NGS and 1000 genome project frequency estimates. Of the ~170 "novel" variants detected in the genes, seven SNPs, all of which were present in multiple sample pools, were selected for validation by Sanger sequencing. Two SNPs were successfully validated by this method, and shown to be genuine variants, while five failed validation. These spurious SNP calls occurred as a result of the presence of small indels and mononucleotide repeats, indicating such features should be regarded with caution, and validation via an independent method is important for NGS variant calls.