Typesafe Coordinate Systems in High-Throughput Sequencing Applications

원문 초록

High-throughput sequencing file formats and tools encode coordinate intervals with respect to a reference sequence in at least four distinct, incompatible ways. Integrating data from and moving data between different formats has the potential to introduce subtle off-by-one errors. Here, we introduce the notion of typesafe coordinates: coordinate intervals are not only an integer pair, but members of a type class comprising four types: the Cartesian product of a zero or one basis, and an open or closed interval end. By leveraging the type system of statically and strongly-typed, compiled languages we can provide static guarantees that an entire class of error is eliminated. We provide a reference implementation in D as part of a larger work (dhtslib), and proofs of concept in Rust, OCaml, and Python. Exploratory implementations are available at https://github.com/blachlylab/typesafe-coordinates.