Come and visit us at AGTA13 October 2014
Come along to learn about some of our recent projects we've been working on, including Genotyping-by-Sequencing and our trial of Oxford Nanopore Technologies MinION or visit one of our team members at booth #2 to discuss your next project.
Bioinformatics approaches, challenges and performance of diversity profiling on three next generation sequencing platforms
Poster #7 (Monday 13th October)
Naga Kasinadhuni and Lavinia Gordon
Australian Genome Research Facility
High-throughput sequencing is reshaping the landscape of microbial studies, revealing thehidden secrets of the “uncultured microbial world”. Diversity profiling has enhanced the discovery and profiling of microbial species that cannot be isolated into a pure culture, by sequencing molecular markers such as the conserved and hypervariable regions of ribosomal RNA. Most samples consist of an extremely heterogeneous microbial community and may contain thousands of species that vary significantly in proportion from one species to the next. Sequencing these samples using high-throughput platforms outputs an enormous amount of data, ranging from 105 to 109 reads per run, with each platform producing different read lengths, quality, error rates and data formats. Despite this abundance of data, deciphering some meaningful information presents a serious challenge to bioinformaticians. From QC to classifying the sequenced amplicons, a series of various complex computation methods are required. The Australian Genome Research Facility (AGRF) currently operates a microbial diversity profiling service targeting various regions of 16S, ITS and 18S microbial genes. To compare and evaluate the accuracy and reproducibility of various platforms, AGRF pooled both artificial control samples and environmental samples, and sequences using GSFLX, MiSeq and Ion Torrent NGS platforms. The resulting data was analysed and classified at different taxonomic levels using various bioinformatics pipelines. Here we present the pipelines and results comparing the three NGS platforms
Efficient Genotyping-by-Sequencing Enables Construction of the First Linkage Map for the Non-Model Legume Forage Species, Bituminaria bituminosa var. albomarginata (tedera)
Poster #38 (Monday 13th October)
Dr Jafar Sheikh Jabbari1, Dr Rust Turakulov1, Dr Kirby Siemering1, Mr Matthew Tinning1, Ms Maria Pazos-Navarro2, Dr Daniel Real3, Dr Matthew N. Nelson4
1Australian Genome Research Facility, 2Centre for Plant Breeding and Genetics, The University of Western Australia, 3Department of Agriculture and Food Western Australia, 4School of Plant Biology, and UWA Institute of Agriculture, The University of Western Australia
Single-nucleotide-polymorphisms (SNPs) identified using genotyping-by-sequencing (GBS) methods are increasingly used in many fields of genetics and population genomics. These methods differ in nucleotide diversity during initial sequencing cycles that affects quality and quantity of reads, number of sampled digestion fragments (tags) and sequencing depth required for confident genotype calls. Here we introduce an improved GBS method based on size-selection of double digested adapter-ligated DNA which increases sequencing yield and quality by producing high diversity libraries. We applied this technique for genotyping a subset of an F2 population from Bituminaria bituminosa var. albomarginata (tedera), an emerging drought-tolerant forage legume species. By sequencing a barcoded library prepared from 45 progeny and parents in one lane of HiSeq we obtained over 150 M SE reads which were analysed using Stacks pipeline resulting in 845 high-quality SNP markers scored in all samples. Linkage mapping identified 11 linkage groups, close to the haploid chromosome number of this species (n=10). Analysis of synteny via BLASTn between the genomes of tedera and its close relative soybean using tedera GBS sequence tags revealed the presence of 264 (31.2%) GBS marker tags with one or more significant (<1e-10) matches in the soybean genome and that marker order was conserved in large chromosomal blocks between the two genomes. This finding, along with application of the method to the remaining progeny, will facilitate the transfer of genomic information between the well-resourced model genome of soybean and the ‘orphan’ genome of tedera.
Stranded versus non-stranded RNA-seq
Poster #94 (Tuesday 14th October)
Dr Sonika Tyagi
Australian Genome Research Facility
RNA-seq has proven to be a powerful technique for transcriptome analysis and has changed the way in which transcription can be viewed. RNA-seq can capture both sequence and abundance of the RNA species in a transcriptome, which makes it an attractive approach to study structure of the transcripts, splice variants and transcriptional activities in non-coding, intergenic or untranslated regions. Following recent proliferation of RNA-seq sample preparation protocols, there is little known about effect of various library protocols on the downstream analysis is incomplete. In this paper we will discuss some of the common library preparation biases their effects on downstream analysis. We will be specifically comparing stranded and unstranded RNA-seq library preparation methods with respect to whole transcriptome analysis.
Pore Performance? Benchmarking the Oxford Nanopore Technologies MinION for Metagenomics and contig assembly.
Wednesday 15th October 3.15pm
Speaker - Dr Ken McGrath
Australian Genome Research Facility
The Oxford Nanopore Technologies (ONT) MinION is a revolutionary new device capable of directly sequencing single DNA molecules via translocation through a protein pore. The Australian Genome Research Facility (AGRF) was chosen to take part in the MinION "early access" sequencing program, which allowed us to receive and trial the platform prior to general release. We have performed several sequencing runs on this device, including whole-genome shotgun sequencing, as well as targeted longamplicon sequencing. These initial runs generated between 1000-5000 reads events per 6hr run, with read lengths ranging from 5bp to over 42,000bp. Analysis of these reads showed a high error rate when compared to the target sequence with a particular tendency to form gaps (both as insertions and deletions).Correct and incorrect calls tended to form clusters along the reads, suggesting the reads can enter an "error-prone state" during certain sections or phases of the read. Both the systematic and random nature of these errors was also assessed.
We assessed the utility of the MinION platform in 2 applications where long sequencing reads are particularly useful - microbial community profiling and de novo assembly. We trialled the platform to identify bacteria from a known artificial microbial community based on sequencing the 16S region, and compared these results with alternate methods available on established sequencing platforms such as the Roche 454 and Illumina MiSeq. We also assessed the utility of MinION reads for scaffolding contigs in the de novo assembly of short-read sequencing data sets.
This work highlights the strengths and weaknesses of this emerging technology, and provides an insight into what the future of single-molecule sequencing holds.