biohazard-0.6.5: bioinformatics support library

Safe HaskellNone




data GenoCallBlock Source

File format for genotype calls.

To output a container file, we need to convert calls into a stream of sensible objects. To cut down on redundancy, the object will have a header that names the reference sequence and the start, followed by calls. The calls themselves have contiguous coordinates, we start a new block if we have to skip; we also start a new block when we feel the current one is getting too large.

compact_likelihoods :: Vector Prob -> Vector Mini Source

Storing likelihoods: we take the natural logarithm (GL values are already in a log scale) and convert to minifloat 0.4.4 representation. Range and precision should be plenty.

getRefseqs :: AvroMeta -> Refs Source

Reconstructs the list of reference sequences from Avro metadata. If a type named Refseq is defined in the schema and is an enum, it defines the symbol table, otherwise an empty list is returned. If biohazard.refseq_length exists, and is an array, it's elements are interpreted as the lengths in order, otherwise the lengths are set to zero.