For an overview of the design principles and use of Bioconductor sequence classes, see Lawrence et al., 2013, Software for Computing and Annotating Genomic Ranges. PLoS Comput Biol 9(8): doi:10.1371/journal.pcbi.1003118

For an overview of select high-throughput sequence packages in Bioconductor, see Intermediate Sequence Analysis 2013 section 3.3.

Ranges

IRanges & GenomicRanges

Classes

• IRanges(), GRanges()
  • Metadata on the object as a whole (metadata()), and on individual elements (e.g., mcols())
  • Behavior inhertted from a base class Vector(), e.g., length(), subset, etc.
• *List(), e.g., IntegerList(), GRangesList()
  • Behave like lists where elements are restricted to be of a common type, e.g., IntegerList() is a list where all elements are integer vectors
  • Implemented as a single instance, e.g., integer() with partitioning vector.
  • Very effective representation for many operations
• DataFrame()
• Rle()
  • Especially useful in ChIP and other regulator contexts, in QA, and in visualization

Methods

• See Intermediate Sequence Analysis 2013 section 6.1.

Users

• Used extensively in high-throughput sequence packages
• Example: GenomicFeatures and related TranscriptDb packages representing UCSC and other genome annotation tracks.

Strings and sequences

Biostrings

• DNAString(), DNAStringSet() classes
  • Derived from XString(), XStringSet classes.
• Extensive methods, see Intermediate Sequence Analysis 2013 section 5.1

Users

• ShortRead uses XStringSet as basis for coordinating short reads and quality scores.
• BSgenome uses DNAString to represent whole genome sequences.

Data containers

GenomicRanges

• GAlignments(), GAlignmentsList(), GAlignmentPairs()
• SummarizedExperiment()

VariantAnnotation

• VCF()

ShortRead – FASTQ files

• ShortReadQ() – Reads and their quality scores

I/O

• rtracklayer

  • Data input from common genomics formats, e.g., bed, wig, gtf, gff, bw
  • Result is usually a GenomicRanges-derived class
  • Single interface, import(), so complexity hidden from user
  • Can also drive (update and retrieve) a UCSC genome browser session

• VariantAnnotation readVcf(), filterVcf(). Manage large data by::

  • Project fields of interest for input with ScanVcfParam().
  • Selecting ranges of interest for input from indexed Tabix files with ScanVcfParam()
  • For specialized uses select INFO and GENO fields for input with readInfo(), readGeno()
  • Iterate through large files using TabixFile(<...>, yieldSize=10000) and a paradigm like

  tbx <- open(TabixFile(fl, yieldSize=10000))
  repeat({
      vcf <- readVcf(tbx, "hg19")    ## up to 10000 records
      if (length(vcf) == 0)
          break                      ## all done
      ## do work
  }
  close(tbx)
  

  • Filter large files to small files using filterVcf()

• Rsamtools BamFile() and TabixFile() to open and iterate through BAM and Tabix files

  • Strategies like those in VariantAnnotation to manage large data: restrict fields and ranges of interest with ScanBamParam(); iterate through large files using yieldSize argument of BamFile().

• Rsamtools

• readGAlignmentsFromBam(), readGAlignmentsListFromBam()

• ShortRead FastqStreamer(), FastqSampler(), readFastq()

  • Examples of reference classes
  • Iterate with yield() on an instance created by FastqStreamer()
  • Sample randomly from an entire file, e.g., for QA purposes, with yield() onan instance created with FastqSampler()

Approximate data class hierarchy

Annotated
   o metadata
-- Vector
   o many methods (showMethods(class="Vector", where=search()))
   -- Rle
   -- List
      -- SimpleList
         -- DataFrame
         -- Simple*List, e.g., SimpleNumericList
      -- CompressedList (IRanges package)
         -- Compressed*List, e.g., CompressedNumericList
         -- Ranges
            -- IRanges
      -- ... *StringSet, e.g., DNAStringSet
   -- GenomicRanges
      -- GRanges (GenomicRanges package)
   -- ... *String, e.g., DNAString (Biostrings package)
      o transcribe, reverseComplement, pairwiseAligment
SummarizedExperiment (GenomicRanges package)
-- VCF (VariantAnnotation package; readVcf)
ShortReadQ