BioPerl: MUMmer

1
BioPerl MUMmer

• Jason Switzer
• Joshua Wu
• Aimee Seufer

2
Agenda

• Background
• BioPerl
• Mummer
• Algorithm/Data Structure
• Suffix Trees (implicit vs explicit)
• Examples
• Limitations for BioPerl
• BioAlignIOmummer

3
What is BioPerl

• A project (developed by volunteer engineers)
• Goal collect computational methods routinely
  used in bioinformatics
• Tools for computational molecular biology
• bioinformatics toolkit for format conversion,
  report processing, data manipulation, sequence
  analysis, batch processing and more
• Open source
• http//bioperl.org/
• Collection of modules (1450)

4
Example
5
Example
6
What is MUMmer

• MUMmer maximum unique exact match
• Definition it is a suffix tree algorithm
  designed to find maximal exact matches of some
  minimum length between two input sequences.

7
Achievements

• Suffix tree a very efficient data structure
• constructed and searched in linear time
• ideal for large scale pattern matching
• Memory usage dependent only on reference sequence
• Finds maximal unique solution to dataset
• How efficient? find all 20 base pair maximal
  exact matches between 2 5 million base pair
  bacterial genomes in 20 seconds, using 90 MB of
  RAM, on a typical 1.7 GHz Linux desktop computer

8
Why Suffix Trees?

• "Suffix trees are widely used in the computer
  field... Recent improvements in the method have
  cut the memory requirement to 17 bytes per
  letter, which brings the method to the verge of
  practicality for bioinformatics applications"
  -- Nat Goodman (Genome Technology).

9
Introduction

• Any string of length m can be degenerated into m
  suffixes, and these suffixes can be stored in a
  suffix tree.
• Setup time O(m) (m is length of string)
• searching time O(n) (n is length of pattern)

10
(No Transcript)
11
(No Transcript)
12
Sample input Homo Sapien

• cagctcctgagactgctggcatgaaggggagccgtgccctcctgctggtg
  gccctcaccctgttctgcatctgccggatggccacaggggaggacaacga
  tgagtttttcatggacttcctgcaaacactactggtggggaccccagagg
  agctctatgaggggaccttgggcaagtacaatgtcaacgaagatgccaag
  gcagcaatgactgaactcaagtcctgcagagatggcctgcagccaatgca
  caaggcggagctggtcaagctgctggtgcaagtgctgggcagtcaggacg
  gtgcctaagtggacctcagacatggctcagccataggacctgccacacaa
  gcagccgtggacacaacgcccactaccacctcccacatggaaatgtatcc
  tcaaaccgtttaatcaataa

13
Sample result
14
Sample input 2 plants

• EARPIVVGPPPPLSGGLPGTENSDQARDGTLPYTKDRFYLQPLPPTEAAQ
  RAKVSASEILNVKQFIDRKAWPSLQNDLRLRASYLRYDLKTVISAKPKDE
  KKSLQELTSKLFSSIDNLDHAAKIKSPTEAEKYYGQTVSNINEVLAKLG

15
Sample output
16
Comparisons Homo Sapiens
17
Chicken
18
Sample Input Chicken

• RVKRVWPLVIRTVIAGYNLYRAIKKK

19
Chicken
20
Investigation

• Explicit suffix trees require more space than
  implicit suffix trees in real data.
• Explicit trees should be used for smaller use of
  storage

21
Limitations on development

• No unified output format for all tools
• Tools available mummer, repeat-match,
  exact-tandems, gaps, mgaps, nucmer, promer,
  run-mummer1, run-mummer3, show-aligns,
  show-coords, show-snps, show-tiling
• Variable command line options
• Poor documentation
• Not user-friendly

22
Difficulties With BioPerl

• Extensive Framework
• everything from IO utilities to BLAST
• Aging Codebase
• lots of copy-and-paste
• old coding techniques data
• Few Developers
• few core developers maintaining the toolkit
• few people understand the
• Uncommon Perl Practices
• much derision over practices such as Tie handles
  and AUTOLOAD

23
Common BioPerl Objects

• BioSeqIO
• reads/writes sequence files (e.g. genbank)
• fully symmetric converter (between various
  formats)
• lots of documentation
• BioSeq
• store various sequences (BioRichSeq)
• used primarily with BioSeqIO
• BioAlignIO
• reads/writes alignment files (fasta)
• not fully symmetric (between formats)
• significantly less documentation
• BioLocatableSeq
• stores locatable sequence data (within another
  sequence)
• used primarily with BioAlignIO

24
Example - Input Data
25
Example - Code
26
THANK YOU FOR LISTENING