GSOC Project Ideas 2015
From GMOD
There are plenty of challenging and interesting project ideas this year. These projects include a broad set of skills, technologies and domains, such as GUIs, database integration and algorithms.
Students are also encouraged to propose their own ideas related to our projects. If you have strong computer skills and have an interest in biology or bioinformatics, you should definitely apply! Do not hesitate to propose your own project idea: some of the best applications we see are by students that go this route. 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.
- Project Idea Name
- Brief explanation: Brief description of the idea, including any relevant links, etc.
- Expected results: describe the outcome of the project idea.
- Knowledge prerequisites: programming language(s) to be used, plus any other particular computer science skills needed.
- Skill level: Basic, Medium or Advanced.
- Mentors: name + contact details of the lead mentor, name + contact details of backup mentor.
Here is a list of the proposed project ideas for 2015:
- Project Idea 1: Using an interpreted language to develop bioinformatics workflows
- Brief explanation: SeqWare is a bioinformatics workflow engine that can be used to chain together the analysis of big data in genomics and bioinformatics. The current workflow language is Java, which is rather verbose.
- Expected results: Use Groovy to hide the current rather verbose Java workflow language. Using an interpreted language also enables for rapid prototyping of workflows. The goal of this is to make scripting SeqWare feel more like shell scripting. This is a similar effort to the GATK team’s Queue, but this would leverage SeqWare. Prototype: https://github.com/larsgt/stimpy
- Knowledge prerequisites: Java, Groovy, git
- Skill level: Medium
- Mentors: Lars Jorgensen, Morgan Taschuk, Pipeline team
- Project Idea 2: Write a Foreign Data Wrapper for Postgres and BAM/VCF
- Brief explanation: SQL is a powerful language that makes querying structured data very straightforward, and genomics produces several types of structured data. Big data from genomics usually comes in two parts: the results, stored in files, and the metadata that describe the results, usually stored in databases. For example, VCF files describe a variant in particular cancer-causing gene, and the metadata will describe what the sample was, where it came from, how it was processed, etc. We would like to use SQL to query both results and metadata together.
- Expected results: Develop a Foreign Data Wrapper for BAM and VCF in order to query alignment and variant information. There is an existing Foreign Data Wrapper for TSV files. This should make VCF and SAM fairly straight forward. Accessing BAM files would be slightly more involved. This could provide a good example of making queries against BAM data. Info: http://www.postgresql.org/docs/9.1/static/fdwhandler.html and http://www.depesz.com/2011/03/14/waiting-for-9-1-foreign-data-wrapper/.
- Knowledge prerequisites: PostgreSQL
- Skill level: Advanced
- Mentors: Lars Jorgensen
- Project Idea 3: Implement a FUSE interface to BAM/CRAM
- Brief explanation: Storage of big data is an ongoing problem that will only get worse. As data moves through a processing pipeline in genomics, the output data is often a lossless conversion of data integrating different information (e.g. FASTQ is a listing of all reads; BAM is an alignment of those reads to a reference but still contains all of the reads from the FASTQ). However, data from earlier in the pipeline is often kept so that the analysis can be repeated with different tools. This results in a duplication of data on the order of gigabytes to terabytes.
- Expected results: Enable a tool to see the same BAM file as either two FASTQs, interleaved FASTQ or whatever format it needs (with the same information). This should be easy to prototype using Python as fuse-python and pysam exists.
- Knowledge prerequisites: C and/or Python, POSIX APIs
- Skill level: Advanced
- Mentor: Lars Jorgensen
- Project Idea 4: Use Galaxy to run SeqWare workflows and process on data
- Brief explanation: SeqWare is a bioinformatics workflow engine that can be used to chain together the analysis of big data in genomics and bioinformatics. SeqWare is currently driven on the command line by skilled users. However, it would be incredibly useful to leverage SeqWare’s robustness and stability for individual non-expert users. Galaxy is a user-friendly mechanism for analysing data that can be used for this task.
- Expected results: There are two potential sub-projects. 1) Adding SeqWare metadata and files as a data source in Galaxy, to enable Galaxy users to use SeqWare data, and 2) Launching and monitoring SeqWare workflows with Galaxy.
- Knowledge prerequisites: Galaxy, Java, web services, PostgreSQL
- Skill level: Medium
- Mentor: Morgan Taschuk
- Project Idea 5: Barcode scanner using phone or tablet to drive LIMS
- Brief explanation: In a typical genomics lab, the Laboratory Information Management System (LIMS) is required to keep track of lot of people, equipment and samples as they interact. A typical LIMS requires a desktop computer and a lot of drop down menus in order to fulfill this task, which takes the technician away from the bench and introduces the potential for error. Large sequencing labs use barcodes instead. Barcode readers are prohibitively expensive for smaller labs. Cameras on phones are getting quite good, so it should be fairly easy to drive the barcode reading from a mobile device. This would be a low cost way for smaller labs to use barcoding in the lab workflows. Barcode reading library: https://github.com/zxing/zxing.
- Expected results: A mobile LIMS application that stores a particular lab workflow and prompts the user to scan barcodes when they reach a particular step in the workflow. It would also be able to send information back to the central LIMS servers.
- Knowledge prerequisites: iOS or Android development, web services, interface design
- Mentor: Lars Jorgensen, Timothy Beck and Tony DeBat
- Project Idea 6: iPython notebook on top of our infrastructure
- Brief explanation: iPython notebook is a powerful tool. It enables reproducible science as people can share their work. It would be interesting to see how iPython notebook and SeqWare could interact. It would also be useful for OICR’s users if they could query our and other metadata using Python or R.
- Expected result: A python library that can be used to query SeqWare’s metadata through their RESTful web service.
- Knowledge prerequisites: Python, web services
- Skill level: Basic
- Mentor: Timothy Beck, Lawrence Heisler, Yogi Sundaravadanam
- Project Idea 7: Use Galaxy to run Reactome analysis and processes on genomic data
- Brief explanation: Reactome is a free, open-source, curated and peer reviewed pathway database. Our goal is to provide intuitive bioinformatics tools for the visualization, interpretation and analysis of pathway knowledge to support basic research, genome analysis, modeling, systems biology and education. Galaxy is an open, web-based platform for data intensive biomedical research, which allows users to perform, reproduce, and share complete analyses.
- Expected results: There are two potential sub-projects. 1) Adding Reactome as a data resource in Galaxy, to enable Galaxy users to use Reactome reaction and pathway annotation data, and 2) Performing identifier mapping and over-representation analysis workflows from Reactome in Galaxy. Reactome Github: https://github.com/reactome/
- Knowledge prerequisites: Galaxy, Java, web services
- Skill level: Medium
- Mentor: Joel Weiser
- Project Idea 8: Biological Graph Visualization
- Brief explanation: Tripal (http://tripal.info) is an open-source suite of Drupal modules that allows a scientific research community to more easily setup and manage a data repository for genomic, genetic and related biological data. It provides data pages, data mining tools and visualizations. Tripal is used or in development by 25 different genome database websites, and is developed by an international group. A Tripal module currently exists for importing, searching and visualizing graph data that models the "network" of interactions of various components of a biological system. However, the module is not complete and requires improvements to the visualizations. The goal of this project would be to complete the remaining work for this module such that it can be shared with others.
- Expected results: Once completed, a Drupal module will freely available for Tripal-based sites to use on their own sites. Thus providing graph visualizations for complex biological systems.
- Knowledge prerequisites: PHP, Drupal, JavaScript, SQL.
- Skill level: Medium
- Mentors: Stephen Ficklin
- Project Idea 9: BioModels AnalysisTools
- Brief explanation: Following the Reactome (http://www.reactome.org) recently launch "Analysis Service" (http://www.reactome.org/AnalysisService/) , the idea is to implement an analysis tool for BioModels (http://www.ebi.ac.uk/biomodels-main/) following a similar approach. The tool has two differentiated parts: (1) data filtering and intermediate data structure creation, (2) high performance analysis tool with a RESTFul Service API to allow programmatic access.
- Expected results: A java core package plus a Spring MVC RESTFul service to be installed in BioModels live site.
- Knowledge prerequisites: Java, SQL, XML, web services, Git.
- Skill level: Medium/Advanced.
- Mentors: Antonio Fabregat, Camille Laibe, Henning Hermjakob
- Project Idea 10: Reactome for Illumina Basespace
- Brief explanation: Reactome (http://www.reactome.org) is a free, open-source, curated and peer reviewed pathway database. One of its goals is to provide intuitive bioinformatics tools for the visualization, interpretation and analysis of pathway knowledge to support basic research, genome analysis, modeling, systems biology and education. Illumina Basespace (http://basespace.illumina.com) is a genomics analysis platform that is directly integrated into the NextSeq, MiSeq, and HiSeq sequencing platforms.
- Expected results: A fully working Illumina Basespace "App" connected to the Reactome Pathway Analysis Service (http://www.reactome.org/AnalysisService/) in order allow Basespace user to perform pathway enrichment and expression analysis agains Reactome Pathways.
- Knowledge prerequisites: JavaScript, json, web services.
- Skill level: Medium.
- Mentors: Antonio Fabregat, Henning Hemjakob.
- Project Idea 11: WebApollo Variant annotation
- Background: WebApollo (http://genomearchitect.org) is an open-source genome browser plugin for conducting manual annotations for genomes inside JBrowse.
- Brief description: This project would add the ability to annotate variants from VCF, multi-sample VCF and GVF files in WebApollo. This would involve creating a better visualization of multi-sample VCF files and creating a server side API for representing these annotations.
- Tools: intermediate javascript, some server side java
- Skill level: Medium
- Mentor: Nathan Dunn, Colin Diesh
- Project Idea 12: WebApollo Multi-scaffold visualization
- Background: WebApollo (http://genomearchitect.org) is an open-source genome browser plugin for conducting manual annotations for genomes inside JBrowse.
- Brief description: Normally genome browsers are capable of displaying one scaffold or sequence at a time, this project would involve displaying (and editing) genes split across multiple contigs at once.
- Tools: advanced javascript, some server side java
- Skill level: Advanced
- Mentor: Nathan Dunn, Colin Diesh
- Project Idea 13: Afra Flexible realtime Export
- Background: Afra (http://afra.sbcs.qmul.ac.uk) is an open-source genome browser plugin for conducting manual annotations for genomes inside JBrowse. For each gene prediction, curations are collected from several users and automatically compared: if all users propose the same changes to a gene model, these changes are considered to be correct. If gene models proposed by different curators disagree, the different gene predictions are shown to several more experience curators who submit their curation in turn. If gene models proposed by the more experienced curators disagree, all predictions are shown to an even more senior curator who makes a final verdict.
- Brief description: Normally gene curation is a finite phase of a genome project. But with Afra (http://afra.sbcs.qmul.ac.uk) we are creating a community of constantly creating curations. We thus need an approach do share the “most up-to-date” version of curations at any point in time. The student should add DAS (distributed Annotation Server) export and appropriate GFF/GTF export functionality that is constantly up-to date - reflecting the latest community contributions.
- Tools: ruby, javascript
- Skill level: Advanced
- Mentor: Anurag Priyam, Yannick Wurm
- Project Idea 14: Implementing filters and filter visualizations for the MGI mouse genome browser using jbrowse
- Brief explanation: Mouse Genome Informatics (MGI, http://www.informatics.jax.org) is the authoritative international bioinformatics database on the laboratory mouse, annotating data and building tools that allow researchers to access comprehensive, integrated information on mouse genes, alleles, phenotypes, disease models, and gene expression in order to facilitate the study of human health and disease. Data within MGI is curated from the primary research literature, or loaded from large-scale projects or other research resources. These data are organized using structured vocabularies (gene function, mutant phenotype descriptions, anatomy) and unique identifiers making all information accessible to computational as well as traditional approaches. At MGI, we use JBrowse as our main sequence and feature browser and are interested in adding to its functionality. We propose adding significant filtering capabilities to JBrowse that are configurable with a visual interface to access these parameters. It would be useful for navigating through dense, complex data such as variants, expression, and transcript features. It will also leverage JBrowse’s ability to download data by allowing users to quickly access downloads with just the transcripts/features they are interested in. This addition would help not only our instance of JBrowse but the JBrowse community.
- Expected results: increased functionality of the browser allowing users to apply customizable filters, view and download filtered results
- Knowledge prerequisites: Javascript, Dojo, Git, some perl, and a basic familiarity with biological data.
- Skill level: Medium
- Mentors: Paul Hale (primary), Joanne Berghout
- Project Idea 15: Intermine for Reactome: a data mining resource for metabolic pathways
- Brief explanation: Biological sciences have entered the the big data era. Tools and techniques are emerging to enable the extraction of patterns and meaning from the growing accumulation of biological data. Data warehousing of biological databases is an integrative approach to making complex information from a variety of sources more easily accessible for data mining and analysis. Reactome is a free, open-source, curated and peer reviewed database of metabolic pathways and their components. Our goal is to provide intuitive bioinformatics tools for the visualization, interpretation and analysis of pathway knowledge to support basic research, genome analysis, modeling, systems biology and education. InterMine is an open source data warehouse built specifically for the integration and analysis of complex biological data. The idea here is to create a new resource that exposes the Reactome Knowledgebase to the well-developed data warehousing tools and interfaces provided by Intermine.
- Expected results: An Intermine implementation for Reactome, including new web interfaces and web services APIs
- Knowledge prerequisites: Java or Perl, XML, relational databases (mySQL, PosgGreSQL)
- Skill level: Medium
- Mentor: Sheldon McKay
