We first mapped the sequences allowing no mismatches (-n 0, seeding disabled) and kept uniquely mapping probes with a minimum mapping quality (MAPQ) of 10 (–q 10). These sequences were then mapped again, this time allowing one mismatch (-n 1). Again, only uniquely mapping probes with MAPQ > 10 were retained, resulting in a total of N = 37,740 probes. We further removed all probes that overlapped with any variant with a minor allele frequency greater than 0.05 in the main imputed dataset (Nlines = 858). Remaining probes were annotated with Gencode version 19 gene annotations 56 and only probes mapping uniquely to a single gene were kept (final probeset Nprobes = 25,604, representing 17,116 unique genes of which 14,569 are protein coding.
