XSQ: A New Binary Output Format of 5500/5500XL Systems

摘要

The XSQ (eXtensible SeQuence) format was designed to accommodate different run types (standard SOLiD sequencing and exact chemistry call (ECC) runs), simplify indexing samples workflow and support new data types such as ECC in the new 5500 sequencing instrument; these data cannot be stored in any existing formats (e.g. csfasta & QV.qual) as there are more than one call per position. Other problems with existing formats are the file size, I/O demand, and lack of pairing of reads. The XSQ format uses information packing to reduce the file size by 60%, resulting in reduced storage needs and reduced transfer times. It also has integrated pairing of reads, so mapping and pairing can happen together much more quickly, rather than having the two reads of a pair mapped separately and then merged subsequently. The new XSQ file allows users to perform very efficient indexing reassignment and such reassignment only introduces minimal impact to downstream analysis instead of reanalyzing all indexing samples. The hierarchical data format also provides a basic level of partitioning and indexing within the file so that subsets of the data can be retrieved without reading through the entire file.