First, we used SEQMINER to query tabix‐indexed files of summary association statistics (as described in section Performance for Annotating Summary Association Statistics). For the file containing single variant association statistics for 39 million variants, SEQMINER took 0.3 sec to retrieve summary association statistics using the function seqminer::rvmeta.readDatabyRange. Not surprisingly, reading the entire file into R (using either data.table::fread or read.table) and performing in‐memory query took considerably longer and required a much larger memory footprint (Table 1). While SEQMINER's main advantage lies in its capability to randomly access files of summary association statistics, and it is often not necessary to read the entire file into the memory, we also compared the speed of SEQMINER in reading the entire file as a benchmark. Using SEQMINER's tabix.read.table function to read the entire file into memory in R took 7.5 min, which is ∼4% faster than R's read.table command.
