-

Google Summer of Code 2014 Ideas

From Open Bioinformatics Foundation
Revision as of 18:26, 14 February 2014 by EricTalevich (talk) (→‎Lightweight BioPerl modules: Jason says he won't have time this year)
Jump to: navigation, search

The details of each of our project ideas are listed below, including potential mentors. Interested mentors and students should subscribe to the OBF/GSoC mailing list and announce their interest.

See the main OBF Google Summer of Code page for more information about the GSoC program and additional ways to get in touch with us.


Cross-project ideas

BioInterchange: Convert and Exchange Biological File Formants using RESTful web service

Rationale
BioInterchange Interchange data using the Resource Description Framework (RDF) and let BioInterchange automagically create RDF triples from your TSV, XML, GFF3, GVF, Newick and other files common in Bioinformatics. BioInterchange helps you transform your data sets into linked data for sharing and data integration via command line, web-service, or API. BioInterchange was conceived and designed during NBDC/DBCLS's BioHackathon 2012. Architecture and RDF serialization implementations were provided by Joachim Baran, Geraint Duck provided JSON and XML deserialization implementations and contributed to architecture decisions, guidance on ontology use and applications were given by Kevin B. Cohen and Michel Dumontier, where Michel brought forward and extended the Semanticscience Integrated Ontology (SIO). Jin-Dong Kim helped to define ontology relationships for RDFizing DBCLS' PubAnnotation category annotations. The main idea is to have a central service with can be used as a validator and as interchange service for different languages.
Approach
The project will identify the most common and used file formants for all the currently used language under OBF and will design a RESTful API and will project an implementation for all the supported languages. BioInterchange was developed with Ruby but the scope of the project is to have an agnostic system which let use implement a converter using the best language for that functionality. It expected to have a high traffic for the service so an appropriate refactoring or reimplementation using parallel techniques or languages devoted to parallel programming would be possible.
Difficulty and needed skills
The project is mid / high difficulty, aimed at talented students. Previous knowledge of Ruby or other scripting language is preferred and flexibility in learning other languages is requireed.
The project requires
Knowledge of advanced programming languages and meta-programming and some concept in parallelizing and web services design.
Mentors
Raoul J.P. Bonnal, Francesco Strozzi, Toshiaki Katayama, Joachim Baran

Language APIs for the Systems Biology Markup Language (SBML) through the JVM

Rationale
The standard Java implementation of SBML, JSBML, is used as a parser for various Java-based systems biology applications. This fulfills one niche, but the versatility of the JVM can be utilized to employ JSBML as a parser for systems biology applications that are written in other languages. Also, JSBML undergoes an active community effort to be up-to-date with current SBML standards.
Approach
This project will aim to present language APIs for languages that may want to employ the SBML structure without building a parser from scratch. Matlab, Mathematica, and Python APIs will be the focus for this project.
Languages and skill
Java, optional: Matlab, Python, (other language)
Mentors
Andreas Dräger, Alex Thomas

BioPerl

BioPerl logo tiny.jpg

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 use 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 II)

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 BioPerl modules

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

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

Rationale
Although progress was made in the 2010 project "Alignment Subsystem Refactoring", continued refinement of AssemblyIO is still needed.
Approach
SAM or ACE files once imported should have similar handles and/or methods.
Difficulty and needed skills
Medium. Proficiency in Perl, familiarity with assembly tools.
Mentors
To be determined.

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.

BioJava

Biojava logo tiny.jpg

For GSoC 2014, BioJava is partnering with the Systems Biology Markup Language (SMBL) team to bring enhancements to JSBML, the standard Java implementation of SBML, and bring SBML features to other Java-based systems biology software. See the SMBL website for more ideas from the SBML team.

Add support for Schema-based validation of SBML

Rationale
SBML files need to be validated carefully to ensure that they conform to the specification. Currently, the most complete implementation of SBML validation is embodied in libSBML, although the rules of SBML validity are defined in the SBML specification documents. It is possible to validate SBML from JSBML using either the Online SBML Validator or a Java package we provide for calling libSBML locally (i.e., without a network connection) but we want to move toward capturing all of the SBML's validity rules in schema languages.
Approach
Capture all of the SBML's validity rules in schema languages such as RELAX NG and Schematron, then have both libSBML and JSBML (and any other SBML-using system) use schema validation engines instead of hardcoded validation. This will be especially important as more SBML Level 3 packages become implemented. We have already made great strides in defining RELAX NG schemas for SBML Level 3, but we need to work on providing the hooks in JSBML to using those schemas for validating SBML files.
Languages and skills
Java, XML, RELAX NG, Schematron, SBML
Mentors
Sarah Keating, Andreas Dräger

Redesign the implementation of mathematical formulas in JSBML

Rationale
JSBML uses the concept of abstract syntax trees to work with mathematical expressions. At the moment, all different kinds of formulas are implemented in one complex class.
Approach
This project should implement a math package for JSBML, in which all different kinds of tree nodes that can occur in formulas (e.g., real numbers or algebraic symbols such as 'plus' or 'minus') would be represented with an own, specialized class. In this way, the handling of formulas would be much more straightforward and even more efficient.
Difficulty and skills
Medium; proficient in Java
Mentors
Andreas Dräger, Alex Thomas, Sarah Keating

Implement support for the SBML Multistate/Multicomponent Species package

Rationale
One of the many packages for SBML Level 3 is Multistate and multicomponent species. This packages define constructs for models and modelers to represent biochemical species that have internal structure or state properties. These may involve molecules that have multiple potential states, such as a protein that may be covalently modified, and molecules that combine to form heterogeneous complexes located among multiple compartments.
Approach
The JSBML team has already started implementation of the multi package, but more needs to be done.
Languages and skills
Java, some exposure to biochemistry
Mentors
Nicolas Rodriguez, Nicolas Le Novère

Improve the plugin interface for CellDesigner

Rationale
One of the most frequently used programs in computational systems biology is CellDesigner. JSBML provides an interface that facilitates the development of plugins for this program. This interface has recently been revised and improved.
Approach
Test cases and plugins for CellDesigner are to be implemented in order to make use of it and ensure its correct behavior. It is, for instance, possible to use CellDesigner's complex canvas user interface to create or manipulate biochemical networks and to conduct numerical computation.
Languages and skills
Java, some basic understanding of visualization algorithms
Mentors
Andreas Dräger

BioPython

Biopython logo tiny.png

Indexing & Lazy-loading Sequence Parsers

Rationale
Bio.SeqIO's indexing offers parsing on demand access to any sequence in a large file (or collection of files on disk) as a biopython:SeqRecord object. This works well when you have many small to medium sized sequences/genomes. However, this is not ideal for large genomes or chromosomes where only a sub-region may be needed. A lazy-loading parser would delay reading the record until requested. For example, if region record[3000:4000] is requested, then only those 1000 bases need to be loaded from disk into memory, plus any features in that region. This is how Biopython's biopython:BioSQL interface works. Tools like tabix and samtools have demonstrated efficient co-ordinate indexing which could be useful here.
Aside from being used via an index for random access, lazy-loading parsers could be used when iterating over a file as well. This can potentially offer speed ups for tasks where only a fraction of the data is used. For example, if calculating the GC content of a collection of genomes from GenBank, using Bio.SeqIO.parse(...) would currently needlessly load and parse all the annotation and features. A lazy-parser would only parse the sequence information.
Approach & Goals
Useful features include:
  • Internal indexing of multiple file formats, including FASTA and richly annotated sequence formats like GenBank/EMBL and GTF/GFF/GFF3.
  • Full compatibility with existing SeqIO parsers which load everything into memory as a `SeqRecord` object.
Difficulty and needed skills
Hard. Familiarity with the Biopython's existing sequence parsing essential. Understanding of indexing large files will be vital.
Possible Mentors
Wibowo Arindrarto, Peter Cock, others welcome

BioRuby

BioRuby logo tiny.png

An ultra-fast scalable RESTful API to query large numbers of genomic variations

Rationale
VCF files are the typical output of genome resequencing projects (http://www.1000genomes.org/node/101). They store the information on all the mutations and variations (SNPs and InDels) that are found by comparing the outputs of a NGS platform with a reference genome. These files are not incredibly large (a typical uncompressed VCF file is few gigabytes) but they are full with information on millions of positions in the genome where mutations are found. Large resequencing projects can produce hundreds or thousands of these files, one for each sample sequenced.
Existing tools (such as VCFTools or BCFTools) offer a convenient way to access these files and extract or convert the information present, but are limited in functionalities and speed when more complex queries need to be performed on these data. With existing tools it is very complicated, if not impossibile, to retrive information when working on many VCF files and samples together to compare, for instance, the variations found in 100 samples and extract all the mutations that are present in 50 samples but are not present in the other 50 and so on.
Approach
The project should develop a RESTful API to address the issues described in the rationale and to allow users to manipulate and compare genomics variation information for hundreds of samples. A database engine will be required to store the information and to support the data mining. Unstructured database engines such as noSQL databases or key-values stores can all be valid alternatives to combine high-speed with data flexibility. The decision on the best database engine to be used will be discussed between the student and the mentors and within the OpenBio community. Given the high amount of information that will need to be processed by such an application, scalable and fast languages such as JVM-based languages like Scala or JRuby will be a good choice. The project should also take care of the deploy of such an API, by creating a Ruby gem or a JAR that users can install and use right away with their datasets.
Difficulty and needed skills
The project has an average difficulty and it is aimed at talented students who wants to develop a fast API to address these problems.
The project requires
Knowledge of advanced programming languages. Some experience and knowledge of databases and data mining will help managing the information of VCF files.
Mentors
Francesco Strozzi, Raoul J.P. Bonnal

BioHaskell

Optimizing transalign, a novel, very sensitive alignment method

Rationale
A method and implementation for more sensitive pairwise alignments was recently developed and published (paper is here, and a copy here). The method appears to be the best of its type -- if nothing else, check the SCOP benchmark -- although it’s difficult to construct a good test case using more complex methods (training sets for HMMs and whatnot). The current implementation is in Haskell, and although it works correctly, it is a bit slow, and more problematic, it consumes too much memory (so going multi-threaded, although pretty easy, won’t be of any help).
Approach
The goal is to make this into a more practical tool by reducing resource requirements. A prospective student would either:
  1. Optimize the Haskell program, primarily to reduce memory footprint, and secondarily to make use of multi-CPU systems.
  2. Reimplement the algorithm (or parts of it) in a different language, and achieving the same as above.
Advantages of 1:
  • Already have a working program, and the type system makes it easy to refactor without introducing errors.
  • Haskell supports lots of good multi-threading programming models (like STM)
  • The author of the method, Ketil, knows Haskell pretty well and will mentor.
Disadvantages:
  • Haskell has some good debugging tools, but they tend to work really poorly for large memory (i.e. it takes a long time to generate profiles)
  • Needs somebody with a bit (or a lot) of experience optimizing Haskell, and good knowledge of high-perf libraries (like vector)
Advantages of 2:
  • Easier to get a student with adequate skills.
  • More predictable performance models in other languages.
  • Easier to compile and install for many users.
Disadvantages:
  • Ideally, should know enough Haskell to read and understand the code.
  • Likely needs a co-mentor with knowledge of the language in question.
Difficulty and skills needed
Medium to hard; Haskell proficiency and some knowledge of algorithm development
Mentor
Ketil Malde (ketil@malde.org)


Retrieved from "https://www.open-bio.org/w/index.php?title=Google_Summer_of_Code_2014_Ideas&oldid=3496"