CCLHunter uses 436 SNPs as barcodes to distinguish cell lines. These SNPs were evenly distributed in each chromosome of the human genome (ChrX and ChrY are excluded) and were screened by the following criteria:

Each allele should be located within the CDS regions of recognized coding genes.

Each allele should be recognized as a biallelic variant, meaning that only two variants (including its reference) could appear in this location, as determined by the dbSNP ALFA[1] project's statistics (build id 20201027095038) [2], and its variation type should be a transversion.

The allele frequency of each locus in dbSNP [2] should fall within the range of 0.4 to 0.6.

The allele frequency of each locus in our curated CCLs should also fall within the range of 0.4 to 0.6.

Each locus should not be within the tandem repeat regions identified by the tandem repeat finder with default parameters [3]

Each SNP should not be located within the linkage disequilibrium regions [4, 5]

If two or more alleles are located on the same coding gene, only the one farthest from the CDS boundary will be retained.