paperKB
coga / coga-kb
Help
Sign in

Chunk #65 — Findings — PLINK 2.0 design — Multiple data representations

Source
Second-generation PLINK: rising to the challenge of larger and richer datasets.
Embedded
yes

Text

As discussed earlier, PLINK 1 binary is inadequate in three ways: likelihoods strictly between 0 and 1 cannot be represented, phase information cannot be stored, and variants are limited to two alleles. This can be addressed by representing all calls probabilistically, and introducing a few other extensions. Unfortunately, this would make PLINK 2.0’s representation of PLINK 1-format data so inefficient that it would amount to a serious downgrade from PLINK 1.9 for many purposes.