IRRI Galaxy: Bioinformatics for Rice Scientists
•Download as PPTX, PDF•
2 likes•2,435 views
Ramil Mauleon talk at ISCB-Asia on IRRI GALAXY: bioinformatics for rice scientists
Recommended
Recommended
More Related Content
What's hot
What's hot (20)
Viewers also liked
Viewers also liked (20)
Similar to IRRI Galaxy: Bioinformatics for Rice Scientists
Similar to IRRI Galaxy: Bioinformatics for Rice Scientists (20)
More from GigaScience, BGI Hong Kong
More from GigaScience, BGI Hong Kong (20)
Recently uploaded
Recently uploaded (20)
IRRI Galaxy: Bioinformatics for Rice Scientists
- 1. IRRI Galaxy: bioinformatics for rice scientists Ramil P. Mauleon Scientist – Bioinformatics Specialist TT Chang Genetic Resources Center International Rice Research Institute
- 2. Presented in behalf of my co-authors & the development team @ IRRI Scientists/product/theme leaders • Michael Thomson • Kenneth L. McNally • Hei Leung Laboratory, software team • Venice Margaret Juanillas • Christine Jade Dilla-Ermita
- 3. Outline • Overview of IRRI & it’s research agenda • Bioinformatics activities at IRRI • IRRI Galaxy: current state, future developments
- 4. International Rice Research Institute: part of the Consultative Group on International Agricultural Research CGIAR
- 5. CGIAR - global partnership that unites organizations engaged in research for a food- secure future • International Rice Research • International Center for Institute (IRRI) Agricultural Research in the • Africa Rice Center Dry Areas (ICARDA) • International Center for • International Institute of Tropical Agriculture (CIAT) Tropical Agriculture (IITA) • International Crops Research • International Livestock Institute for the Semi-Arid Research Institute (ILRI) Tropics (ICRISAT) • International Water • International Maize and Management Institute (IWMI) Wheat Improvement Center (CIMMYT) • International Potato Center (CIP)
- 6. INTERNATIONAL RICE RESEARCH INSTITUTE Los Baños, Philippines Mission: Reduce poverty and hunger, Improve the health of rice farmers and consumers, Ensure environmental sustainability Through research, Home of the Green Revolution partnerships Established 1960 www.irri.org Aims to help rice farmers improve the yield and quality of their rice by developing.. •New rice varieties •Rice crop management techniques
- 7. Global Rice Science Partnership : GRiSP • A single strategic and work plan for global rice research • Streamlines current research for development activities of the CGIAR, aligns it with numerous partners, and • Adds new activities of high priority, in areas where science is expected to make significant contributions. IRRI +++
- 8. 6 GRiSP Research Themes (2 are rice – research, per se) 1. Harnessing genetic diversity to chart new productivity, quality, and health horizons 1.1. Ex situ conservation and dissemination of rice germplasm 1.2. Characterizing genetic diversity and creating novel gene pools (SNP genotypes, whole genome sequencing, phenotypes) 1.3. Genes and allelic diversity conferring stress tolerance and enhanced nutrition (candidate genes) 1.4. C4 rice (Converted from C3 photosynthesis) 2. Accelerating the development, delivery, and adoption of improved rice varieties 2.1. Breeding informatics, high-throughput marker applications, and multi-environment testing
- 9. IRGC – the International Rice Genebank Collection World’s largest collection of rice germplasm (located at IRRI) held in trust for the world community and source countries • Over 117,000 accessions from 117 countries • Two cultivated species Oryza sativa Oryza glaberrima • 22 wild species • Relatively few accessions have donated alleles to current, high- yielding varieties • http://www.irri.org/GRC
- 10. Rice is morphologically very diverse
- 11. Structure of O. sativa 45 SSR Loci on 2252 lines. (DARwin5, unwtd NJ, SM coef.) The color represents group assignment for K= 9 with a minimum allele frequency of 0.65 for model-based structure analysis. IRRI CORNELL Rice exhibits deep population structure.
- 12. A high quality reference genome is available HQ BAC-by-BAC Nipponbare (< 1 error in 10K bases) IRGSP 2005 Nature 436:793-800
- 13. Research themes, Bioinformatics & Galaxy • Leveraging the reference genome, datasets are sequencing technology-based o Requires bioinformatics knowledge o Small bioinformatics team at IRRI = • We need to o enable field/bench researchers for bioinformatics o share bioinformatics solutions across GRiSP partners o share solutions with rice research community as a whole • Galaxy bioinformatics workbench (http://galaxyproject.org/) an easy choice
- 14. Galaxy features that fit our needs Open, web-based platform for accessible, reproducible, and transparent computational biomedical research. • Accessible: Users w/o programming experience can easily specify parameters and run tools and workflows. • Reproducible: Galaxy captures info so that any user can repeat and understand a complete computational analysis. • Transparent: Users share and publish analyses via the web and create interactive, web-based documents that describe a complete analysis.
- 15. GRiSP 1.2.1: Rice SNP Consortium for enabling genome- wide association studies • Data from high-density genotyping using 44K, 700k Affymetrix SNP arrays and Illumina Beadstudio, Fluidigm medium density platforms • Bioinformatics needs GRiSP 2.1.3 High- throughput SNP • Genotype data management system: SNP genotyping calling, storage, integration, retrieval, platform for formatting for analysis breeding applications • Analysis: GWAS pipelines, genetic analysis tools (for standard & specialized populations) • Genome browser: integrating published datasets & visualizing
- 16. Our 1st Galaxy: SNP calling workflow at IRRI GenomeStudio + BeadXpress Scan Alchemy plug-in Results ( 384 SNPs) Allele calling with ALCHEMY
- 17. Why ALCHEMY SNP calling • GenomeStudio’s genotype calling algorithm is designed for human applications o does not consider inbred samples or population deficient in heterozygotes • Alchemy : Open source, developed at Cornell University by Mark “Koni” Wright et al. (2010) o addresses the poor performance of the vendor’s software on inbred sample sets o ability to estimate and incorporate inbreeding information on a per sample basis o written in C ; compiles neatly under the GNU/Linux environment
- 18. GRISP 1.2.3: The Rice 3,000 Genomes Project: Sequencing for Crop Improvement Kenneth McNally, Ramil Mauleon, Chengzhi Liang, Ruaraidh Sackville Hamilton, Zhikang Li, Ren Wang, Hongliang Chen, Gengyun Zhang, Hongsheng Liang, Hei Leung, Achim Dobermann, Robert Zeigler CAAS + Many Analysis Partners NIAS Cornell TGAC MIPS Cirad IRD CAS CAAS BGI Academia Sinica MPI KZI EMBRAPA AGI Wageningen CSHL Gramene Plant Onto … Uni Queensland …
- 19. Bioinformatics challenges of the project… • Efficient database system that allows the integration of the genebank information with phenotypic, breeding, genomic, and IPR data for enhanced utilization • Development of toolkits/workbenches to enable gene/genotype->phenotype predictions by research scientists and rice breeders • Make these databases, tools, & analyses results available (& updated) along with the rice gene bank
- 20. Focus of bioinformatics developments in 3k project • Sequence/genotype data management, manipulation system o include primary data visualization (SNPs, genome) • Data analysis workbench o Analysis tools, w/ workflow management o Results visualization (haplotypes, population structures, GWAS results) o Highly efficient sequence/analysis results data storage model & phenotype database
- 21. Objective 1 : Sequence primary analysis • Milestone 1: Construction of new variety group reference genomes for the representative clades o Quick draft genomes: SOAP de novo –based assembly (Assembl, V.J. Ulat - IRRI) • Velvet fails with our dataset (legitimate out-of-memory error, likely due to repeats) o New strategies (adapt/optimize/create algorithms) for high-quality assembly of new references, thru collaborations with partners mentioned before..
- 22. New k-mer size Assembl iteration SOAP denovo Automatically assembly generate •Contig SOAP denovo •Scaffold config files •Gap closer Draft genome •with tiling path Short •multi- reads mapped, unmapped data scaffolds Align scaffolds to QC reference trim/filter (nucmer) (fastx •Bin to toolkit) chromosomes •Segregate per Reference chromosome genome(s) unique, multi-hits
- 23. Objective 1 : Sequence primary analysis (contd) Milestone 2: SNP genotypes construction & diversity analysis: Haplotype structure & local (genome- block) diversity analysis o Main problem: • Number of samples (3,042 varieties) overwhelms existing software & computers (for SNP discovery, a big problem) o One Proposed Solution : PANATI • Koni Wright PhD thesis, Cornell University – Very fast SNP discovery and genotype calling using SW alignment
- 24. PANATI (http://panati.sourceforge.net) • No hard limits on the number of mismatches and in/dels imposed by the algorithm • Designed for and best suited for analysis of population samples with high diversity or for the use of a divergent proxy reference sequence for species which have no adequate reference of their own • Fast execution even when there is high divergence between the sample and the reference sequence • free for academic use
- 25. PANATI technical features • Read lengths of any size o Input can be mixes of different read lengths and single- end or paired-end formats • Flexible trade-offs between speed and memory usage • Multithreaded parallel execution of mapping and alignment scaling in linear performance up to 64 CPUs (higher has not been tested) • Ability to read compressed FASTQ files in bzip2 or gzip formats directly o will automatically use pbzip2 for parallel decompression of pbzip2 compressed files if the program is available
- 26. Objective 1 : Sequence primary analysis (contd) • Milestone 3: Annotation of constructed variety reference genomes, genotypes/haplotypes of the 10k genomes, & diversity analyses results o Intersection of results from various annotation pipelines • RAP pipeline(NIAS , T. Itoh et al) • PASA (TIGR) • Gramene evidence-based method • Maker (GMOD)
- 27. Objective 2 : Build database & visualization tools for the genomes / genotypes / haplotype/diversity analysis results Milestone 1. Building the project genome browser; some issues: o Multiple reference genomes to display & call SNPs from • Per reference view, several at a time • Super (“pan”) genome view o So many varieties to display • Pick & show subsets? Global Display? • Regional/global genome comparisons between varieties
- 28. Option 1: UCSC Genome Browser • Good o Fast even for large datasets o Funded, with large community support base o Nice integration with Galaxy • Pick & choose varieties in Galaxy UCSC gbrowser visualization • Not so good o Painful installation o Steep learning curve (esp. for customizations) o Lack of comparative genome view
- 29. UCSC Browser hosted @ CU, mirror @ IRRI
- 30. Option 2: GMOD Gbrowse • Good o “Comfort zone” genome browser – installation, customization o Simple DB schema (basic install) o Funded, with large community support base o Comparative genome view supported o Integrates with Galaxy (similar to UCSC Gbrowser) • Not so good o Slow for large datasets
- 31. GMOD Gbrowse with draft genome assembly anchored rice reference genome
- 32. Objective 2 : Build database & visualization tools for the genomes / genotypes / haplotype/diversity analysis results (contd) Milestone 2: Build data analysis application tools coupled to the sequence database • Some existing tools (input from collaborating institutes) o EU- transPLANT project: computational infrastructures for plant genomics o Haplophyle @ CIRAD • Build Galaxy for tools developed/adopted by project o Sequence/genotype management o Novel data analysis methods, workflows
- 33. Objective 3 : Genotype - > Phenotype analysis/ breeders’ toolkit • Milestone 1 – Create an integrated phenotype database • Milestone 2 - Association (GWAS) & genetic analysis tools o TASSEL , java web start in IRRI GALAXY o R packages integrated into IRRI GALAXY • R-GENETICS • GAPIT – Buckler, et al., Cornell University • Milestone 3 – The breeders’ toolkit o Major project.. Putting all these tools together in a target user-friendly package o Breeder’s use cases captured as workflows Is GALAXY up to this task?? Will breeders use it??
- 34. IRRI Galaxy: Current status • Deployed in the cloud (Amazon Web Services Large instance – Singapore region) • Streamlined to contain rice-specific tools and genotyping data • NO NGS assembly tools in public site
- 37. Workflows for rice data analysis already available
- 38. IRRI Galaxy Toolshed is under development (1)
- 39. IRRI Galaxy Toolshed is under development (2)
- 40. Share data, import into current analysis (upon publication of studies..)
- 41. Solving the data mining issue for large data/results sets • BIO HDF5 technology (Hierarchical Data Format) - http://www.hdfgroup.org/projects/biohdf/ • Bottom line: o very fast data mining of alignments (SAM/BAM), sequences when the data model/file organization & tools (C APIs & libraries) are used o Pilot ongoing now for 2,000 samples genotype data HDF5 •BAM/SAM C API •Sequence files •SNP data loader queries analyses results •Sequences File system •annotation
- 43. Projects in IRRI Galaxy bioinformatics workbench • SNP data pre-processing & calling (Alchemy, PANATI - M. Wright) • Data format manipulation for downstream analysis tools • Population analysis tools o Structure (Pritchard et al.) o Ade4 R package (Chessel et al.) for Analysis of Molecular Variance • Downstream sequence analysis tools e.g. unique primer design (Triplett et al, Colorado State University, in prep) • Interfaces for SNPs data management & analysis o GWAS: TASSEL (Bradbury et al.), GAPIT o GBS analysis pipeline • Pick & choose data to visualize: Varieties Genome browser
- 44. Summary Bioinformatics and database to Use in Integrate sequence-phenotype data breeding BGI de novo and programs Rice SNP Consortium re-sequencing 700k Affymetrix Initial 5-10X coverage genotyping chip 2000 lines 10,000 GeneBank accessions1 Cultivated + close wild relatives Genebank as a reverse genetics system Phenotyping network Association genetics and 2000+ lines QTL mapping •Select Predict genotype-phenotype accessions relationships at kb resolution based on QTL prediction for targeted Specialized genetic phenotyping of stocks: MAGIC specific traits populations, biparental RILs, CSSL, •Discover novel phenotypes IRRI & GRiSP, CAAS 1 Including publicly accessible germplasm from IRRI, CIRAD, AfricaRice , CIAT and regional collections
- 45. THANKS FROM OUR CUSTOMERS