Google Summer of Code

From BioPerl
Jump to: navigation, search

Contents

Mentor Volunteer List

BioPerl developers who volunteer to act as a mentor to a GSoC student.

Project Ideas for 2014

Please add any new project ideas below. We will post updates here and to the mail list regarding timelines, deadines, etc.

Below are a few example project ideas from the last application period.

NGS-friendly BioPerl code

Rationale 
BioPerl is known to be slow re: any data sets, but particularly when dealing with very large data (e.g. anything related to NGS analysis. Can we make it better? Where should we focus our efforts?
Approach 
Under the supervision of their mentor(s), the GSoC student will:
  • Benchmark bottlenecks that lead to loss in performance for NGS analyses
  • Refactor old classes or develop new optimized code for NGS analysis
Challenges 
This can be a self-contained project, but will require a lot of discussion on what areas to focus on.
Difficulty and needed skills 
easy to hard, depending on student's familiarity with the tools to be used. Student will need:
  • excellent Perl programming skills, including familiarity with NGS datasets
  • knowledge of modern Perl practices.
Mentors 
Chris Fields, others?

Convert BioPerl-DB to DBIx::Class

Rationale 
Bioperl-db (the BioPerl bindings to BioSQL) in essence constitute a self-made ORM, invented at a time when DBIx::Class didn't exist yet. As such, it has some advantages (if you are willing to count overly clever features to be counted in this category), but arguably many more disadvantages, chief among them being the unsustainably small (you could also say non-existent) developer community supporting it, and the fact that DBIx::Class now has existed for years, and is fairly mature. So, rewriting Bioperl-db with a DBIx::Class (or another well-supported generic ORM) would stand to make a considerable impact on our ability to further develop Bioperl's relational storage capabilities, as well as BioSQL itself.
Approach 
Under the supervision of their mentor(s), the GSoC student will:
  • Start working on conversion of BioPerl-DB classes to using DBIx::Class
  • write additional tests and improve documentation as needed
Challenges 
BioPerl-DB is self-contained; this may require looking at the BioSQL schema and determining whether there are specific areas that need the most focus.
Difficulty and needed skills 
easy to hard, depending on student's familiarity with the tools to be used. Student will need:
  • excellent Perl programming skills, including familiarity with:
    • DBIx::Class
Mentors 
Hilmar Lapp, others?

Major BioPerl Reorganization, part 2

Save the monolith!
Rationale 
The initial run at this project had some success, but more work needs to be done. The final goal of this project is to find and break out as many well-defined subsections of BioPerl as possible, releasing them to CPAN along the way.
Approach 
Under the supervision of their mentor(s), the GSoC student will:
  • break current thousand-module monolithic distributions into smaller, more manageable pieces
  • improve characterization of dependencies
  • improve build and testing systems for new distributions
  • write additional tests and improve documentation as needed for the reorganization
Challenges 
BioPerl contains nearly 2000 modules, with very complex relationships between them.
Difficulty and needed skills 
easy to hard, depending on student's familiarity with the tools to be used. Student will need:
  • excellent Perl programming skills, including familiarity with:
    • testing (prove, TAP::Harness)
    • module authoring (Module::Build,Dist::Zilla,PAUSE)
  • good knowledge of command-line text-processing tools like ack, grep, and Perl one-liners.
  • version control systems (BioPerl uses git).
Mentors 
Chris Fields, others?

Perl Run Wrappers for External Programs in a Flash

Rationale 
BioPerl has a long tradition of providing wrapper objects for running external programs and parsing their output, mainly through the distribution called bioperl-run. Wrappers make it relatively easy to process data in highly customizable pipelines with the benefits of BioPerl objects and I/O. They also help to standardize the interfaces to typically idiosyncratic open-source utilities, reducing the burden on the developer. With new bioinformatics tools being released almost daily, however, it can be difficult for the BioPerl regulars to maintain a stable of run wrappers for the latest and greatest tools. Even harder is making the wrapper interfaces themselves conform to a standard API that users can count on.
Possible approaches
  1. Integrate Galaxy's tool configuration file format in a pluggable way for developing a generic wrapper application.
  2. Improve/tighten/extend the Bio::Tools::Run::WrapperBase and Bio::Tools::Run::WrapperBase::CommandExts system for very general run wrappers, making them work robustly with the new Bio::Tools::WrapperMaker module currently under development. The goal will be to get these modules ready for release into the trunk.
  3. Are there any shortcomings to current schemes, such as Galaxy's or EMBOSS's acd format, that could be addressed with a newer schema?

See HOWTO:Wrappers and the above module documentation for more details.

Difficulty and needed skills 
Medium. The student should understand or be willing to work hard at understanding BioPerl object-oriented style. Some familiarity with XML and XML Schema will help in getting up to speed. An interest in playing with new open-source bioinformatics tools, especially those for managing next-generation sequence assembly, would also be valuable.
Mentors 
Mark Jensen, Chris Fields

Lightweight/Lazy BioPerl Classes

Rationale 
Many current BioPerl classes are implemented in a greedy or heavy way, where all information is pulled into memory as objects. For instance, the current Bio::Seq implementation is the primary bottleneck for sequence parsing speed and can take up a ton of memory, particularly with whole-genome information and next-generation sequencing information. Storing the data in memory in a simple data structure and generating the objects lazily could help with speed. Alternatively, storing the data in a persistent manner would also help with memory issues, with the obvious trade-off for speed but having the nice side-benefit of consistent and possibly persistent ways of handling data.
Approach 
Implement a Bio::Seq/Bio::PrimarySeq class (or other commonly-used BioPerl classes) that can deal with very large datasets in a memory-efficient manner. Implement at least one corresponding parser that can either parse records lazily (akin to an XML pull parser) or create lightweight objects. These could be considered two projects but they are interrelated (lightweight objects could have many different backends, including lazy parsing), so development should proceed with this in mind.
Difficulty and needed skills 
medium to hard. Student should have an excellent command of Perl and data structures, experience with persistent storage mechanisms (such as a SQL-based RDBMS, CouchDB, etc), and some familiarity with parsing methodologies.
Prior art 
Jason Stajich has started a SQLite-based lightweight Bio::Tree::Tree implementation on a GitHub branch at the recent GMOD Evolutionary Biology Hackathon at NESCent in Fall 2010.
Mentors 
Chris Fields, Jason Stajich


BioPerl 2.0 (and beyond)

Rationale 
Design or reimplement BioPerl classes without API constraint, using Modern Perl tools or Perl 6.
Approach 
Most BioPerl code is over 6 years old and doesn't take advantage of Modern Perl tools, such as new methods available in Perl 5.10 and 5.12, Moose/MooseX, DBIx::Class, Catalyst, and more. Furthermore, a viable Perl6 implementation, Rakudo, is currently available. This gives us an enormous opportunity to redesign fundamental aspects of BioPerl without the necessity for development hindered by a requirement for backwards compatibility.

Two projects, Biome (Moose-based BioPerl) and BioPerl6 (Perl 6 BioPerl) have already started but are in a very early stage. One could participate in:

  • IO implementations for object iteration, or Perl6 grammars for common formats
  • Redesign of common BioPerl classes
  • etc.

This is an area ripe for new student project ideas. The more focused the better! Discussion is a must, either via IRC or email.

Difficulty 
Project-dependent
Mentors 
Chris Fields, Rob Buels

Bio::Assembly

Continued refinement of AssemblyIO - sam or ace files once imported should have similar handles and/or methods.

Semantic Web Support

Rationale 
There are great development opportunities in information discovery for bioinformatics using semantic web, specially thinking in the implementation of SPARQL queries for a "discoverable bio-cloud".
Approach 
Previous efforts can be adopted and extended, such as resulting code from BioHackathon 3 and the code provided by Expasy. Using the modules of the Semantic Web with Perl community, built around RDF::Trine low-level API. There are two main areas to explore:
  1. Parsers and converters from and to RDF, including IO modules for GenBank, EMBL, several XML specifications, et cetera.
  2. Storage and retrieval of information using SPARQL.
Difficulty and needed skills 
Medium. Familiarity with SeqIO modules and Perl itself. The student should also be familiar with RDF format and the RDF triples concept for Semantic Web.
Mentors 
To be determined. Kjetil Kjernsmo can help mentor students wishing to explore the RDF::Trine direction.

(your idea here)

Please feel very free to propose your own idea. As long as it is relevant to one of our projects, we will give it serious consideration. Creativity and self-motivation are great traits for open source programmers.

Do not hesitate to propose your own project idea: some of the best applications we see are by students that go this route.

Past Projects

2011

Major BioPerl reorganization

Save the monolith!
Rationale 
BioPerl is currently suffering from an overly-monolithic structure, which is becoming unwieldy and contributing to paralysis of the project.
Approach 
Under the supervision of their mentor(s), the GSoC student will:
  • break current thousand-module monolithic distributions into smaller, more manageable pieces
  • improve characterization of dependencies
  • improve build and testing systems for new distributions
  • write additional tests and improve documentation as needed for the reorganization
Challenges 
BioPerl contains nearly 2000 modules, with very complex relationships between them.
Difficulty and needed skills 
easy to hard, depending on student's familiarity with the tools to be used. Student will need:
  • excellent Perl programming skills, including familiarity with:
    • testing (prove, TAP::Harness)
    • module authoring (Module::Build,Dist::Zilla,PAUSE)
  • good knowledge of command-line text-processing tools like ack, grep, and Perl one-liners.
  • version control systems (BioPerl uses git).
Student
Sheena Scroggins
Mentors 
Robert Buels, Chris Fields
Code blog
http://techomics.com/

2010

Alignment Subsystem Refactoring

Rationale 
BioPerl's Bio::Align::AlignI subsystem is quite old and in need of significant refactoring. Furthermore, the Bio::AlignI and Bio::Assembly subsystems need further integration. This is an area ripe for reimplementation to make a more consistent set of modules.
Approach 
see the Align Refactor page for more details.
Difficulty and needed skills 
medium to hard. Excellent command of Perl, familiarity with sequence alignment and alignment tools.
Mentors 
Chris Fields, Mark Jensen
Student 
Jun Yin
Code blog 
http://gsoc2010-junyin.blogspot.com/

2009

As part of NESCent's Phyloinformatics GSoC

2008

As part of NESCent's Phyloinformatics GSoC

Publications

  1. Han MV and Zmasek CM. phyloXML: XML for evolutionary biology and comparative genomics. BMC Bioinformatics. 2009 Oct 27;10:356. DOI:10.1186/1471-2105-10-356 | PubMed ID:19860910 | HubMed [phyloxml1]
Personal tools
Namespaces
Variants
Actions
Main Links
documentation
community
development
Toolbox