Biomedical Engineering Reference
In-Depth Information
are implemented as virtual so as to allow for class extensions, which
provide full support for:
BED format (GenomicRegionBED class);
GFF format (GenomicRegionGFF class);
SAM format (GenomicRegionSAM class).
The virtual methods are redefi ned - when necessary - for each derived
class in order to properly read, print, and update the extra variables of
each format.
Additionally, for each format, we provide simple classes whose goal is
to only read in the corresponding format and immediately convert it into
the REG format. These classes are used when the extra variables of the
input format are not needed for a particular computation, and can save
both time and memory resources. These classes are:
GenomicRegionBEDToREG;
GenomicRegionGFFToREG;
GenomicRegionSAMToREG.
8.4.2 The GenomicRegionSet class
The GenomicRegionSet class implements the notion of a set of genomic
regions that corresponds to an entire input fi le, for example the set of
known genes, or aligned reads from a sequencing experiment. The main
data element stored in this class is an array of instances of the
GenomicRegion class (or any of its derived classes). The input fi le can be
read from the standard input to facilitate pipelined execution, and is not
necessarily loaded fully in memory in order to minimize memory
requirements: essentially, the input fi les are read and processed sequentially
one line at a time, and the data for each line are discarded when no longer
needed for the particular computation. In the next section we show an
example of allocating an instance of this class. The operations of the
genomic_regions command-line tool summarized in Table 8.1 are
implemented as methods in this class (see the User's Manual for details).
￿ ￿ ￿ ￿ ￿
8.4.3 The GenomicRegionSetScanner class
This class helps scan by sliding windows an instance of GenomicRegionSet
and is used by the genomic_scans command-line tool. The main advantage
 
Search WWH ::




Custom Search