Blog @ Illumina
Real scientists. Real commentary.

Live from the Asia-Pacific Scientific Summit, Part Two

by
Brett Kennedy and Sonia Gowan
| May 07, 2014

Day 2 of the Illumina ISS arrived after a night of entertainment and numerous networking discussions from Day 1, I am sure many collaborations were incubated in this fertile environment. In order to provide as wide a range of topics from the NGS menu the day has been split into 3 tracks; Human Disease Research and Clinical to Translational in one room, General, Agricultural and Applied in a second and a technical stream presented by Illumina support experts from sample prep to informatics in a third.

Human Disease Research/Clinical Translation Track

Graham Taylor from the University of Melbourne spoke about the potential for next-generation sequencing (NGS) technology to disrupt the cost of testing for inherited conditions. His current focus is comparing targeted and  exome sequencing approaches for dementia. In Australia there are ~321,000 people suffering from dementia in a population of about 23M (1.4%). With a forecasted increase of 20% to over 400,000 over the next 20 years, this will be an increasingly significant disease burden. To assess the impact of genetic factors on dementia, in consultation with local neurologists developed a custom Nextera Rapid Capture panel (100% CDS design coverage) and this was compared to Nextera Rapid Capture Exome (94.2% design coverage) and TruSight One (92.8% design coverage). They sequenced a Yoruban control trio + 48 samples on HiSeq2500 to 1000x average depth, and found merging forward and reverse reads aided alignment ability and there was 100% concordance between the technologies for SNVs in the known Yoruban controls. While all three Illumina solutions worked well, the main advantage of their custom design was the lower sequencing cost per sample but also the ability to target splice regions in introns. They also found it was possible to extract copy number data from regions of high sequence coverage.

Professor Tai-Gyu Kim discussed HLA typing using MiSeq and a custom approach developed by the Catholic University of Korea. HLA is an interesting locus for sequencing as it has significant implications for stem cell and bone marrow transplantation and has also been implicated in having a role in a number of complex diseases such as rheumatoid arthritis, psoriasis, and narcolepsy based on GWAS data. There are ~258,000 donors in the Korean HLA registry and they were able to develop a high quality, high throughput sequencing test targeting Class I and Class II HLA genes using a nested PCR approach on the MiSeq. Graft-versus-host disease (GVHD) is a serious disease in transplant recipients leading to tissue rejection where 5% of unrelated transplant recipients die (Class IV GVHD). Interestingly, they found little variance in GVHD rates between related and unelated haploidentical (HLA genes) transplants. Looking at a number of cytokines, chemokines, and toll-like receptors in GVHD cases, they found SNPs associated with GVHD and these were different between haploidentical family tissue donors. I look forward to hearing more about this in the future as they continue this line of investigation.

Dr. V. L. Ramprasad from MedGenome Labs presented on Indian genomic diversity. As many people know, India is an incredibly culturally diverse continent and this also appears to be reflected in the genetic structure of its people. Given the very large population, sizeable numbers of people are affected by genetic disease but in a geographically divergent way. For example, various types of the blood disorder thalassemia are carried by 40% of people in the north of India but only 3% in the south. However, 3-4% of the affected population amounts to 30-40M people. Exome studies suggest significant divergence from 1,000 genomes SNP frequencies. Understanding these frequencies is a critical element in assigning pathogenicity to variants. MedGenome is using data from their studies to show a number of HGMD-assigned pathogenic variants are too frequent in the Indian population to be pathogenic based on reported disease frequency. This is an exciting opportunity for population sequencing, and they already plan to sequence 150 genomes to kick start this effort. Good luck!

Single cell analysis, named method of the year for 2013, is a hot topic this year too. Aaron Streets from Peking University presented on a microfluidic device they have developed that enables single cell collection and transcriptome analysis. The reason for this interest is that in a typical sample with multiple cells, current methods deliver the average of activity from the mixture. Single cell methods enable the identification of specific expression profile modes (such as a bimodal distribution amongst cells) that would be hidden using standard analyses. Automated cell collection and RNA extraction enables a reduction in contamination, and reduces systematic variation and human errors. When the team looked at low abundance reads (RPKM=1), they got twice as many reads from their device as using a tube-based method and overall were able to achieve a >90% single molecule detection rate. With this  improved sensitivity and precision, combined with a high capture rate and the ability to perform pre-processing (e.g. drug treatments prior to sequencing) watch for more exciting research to continue to push the envelope on single-cell analysis.

Hepatocellular carcinoma (HCC) is a cancer with a variety of causes including viral infection, alcohol consumption, aflatoxin and obesity. Tatsuhiro Shibata from the Japanese International Cancer Genome Consortium has focused on viral- and alcohol-induced HCC. Using whole-exome sequencing on HiSeq, four unique mutational signatures were seen.

Ethnicity

Japanese

Asians in United States

Caucasian

Mutational Signature

1,3,4

2,3

3,4

 

Interestingly, there was one common signature (no surprises that this was an age-related response) and different signatures depending on ethnicity and environment. Given the wide range of HCC causes, more functional work is needed to identify the drivers of these differences.

Marcel Dinger from the Garvan Institute then presented results from a research program into Intellectual Disability (ID), which has been part of the institute’s overall transition to a clinical testing laboratory. Historically, genetic testing has been very expensive, but with NGS that is changing. ID is also a surprisingly common disease and is found in about 3% of the population with 183,000 children and 420,000 adults affected. In their research into 53 families with an ID child, they identified a probable cause from 20/53 (40%) and suspicious novel variants in 16/53 (30%) using exome sequencing. But in discussing the thinking behind their HiSeq X Ten purchase, Marcel pointed out that exome capture products tend to cover only 85-90% of coding regions and often miss intronic splicing regions, whereas whole genome sequences covers 96-97% including the non-coding regions for which definitive disease linked variants are being increasingly identified. In fact Marcel felt that “WGS is the best exome test you can do” and he believed the diagnostic yield from WGS will be greater than 50%.

Agricultural Genomics Track

New Zealand dairy farmers are extremely fortunate to have had the foresight to partner with scientists at Livestock Improvement Corporation, who are aiding the search for  causative variants underlying commercially important traits.  Matt Littlejohn from New Zealand’s Livestock Improvement Corporation (LIC) gave us a glimpse into the incredibly rich national dairy stock data set; LIC have brd 4 out of 5 cows in New Zealand!  Using the Bovine HD Genotyping BeadChip, they have collected genotype information on >100k animals with ~777k SNPs, plus the phenotypes of ~70k animals with ~10 basic milk traits, ~850 animals with ~50,000 detailed phenotypes (e.g., microarray of liver and fat tissues, milk and blood metabolics). Using an integrated multidimensional approach, LIC precompute a database of all SNPs x all traits and use an algorithm called “Pleiotrapper” to generate a set of scripts and return all significant associations. This way, LIC can identify pleiotrophic effects between physiological phenotypess and gene expression phenotypes and identify causative genes through correlation of eQTL and physiological trait QTL. LIC also performed RNA-Seq on mammary tissue and found transcriptome data can be used to shortcut causative gene identification. Dr. Littlejohn highlighted that the source tissue matters as many eQTL’s are tissue specific and phenotype-rich datasets can reveal novel biological associations. Perhaps surprisingly, ~550 animals have been whole-genome sequenced, finding ~25M variants. Early data indicates that imputation can be used to yield sequence-resolution genotypes in large populations, may yield improved association over chip-based genotypes, and allow rapid identification of causative genes and variants.

Nicotiana benthamiana, a herbaceous relative of tobacco indigenous to Australia, is truly the plant equivalent of the nude mouse. The plant has the rare ability to rapidly express transgenes simply injected into its leaves. In a talk on viruses, genes, genomes and small RNA-regulated plant processes, Peter Waterhouse of the Queensland University of Technology reported the genome and transcriptome sequencing of this ‘wonder plant’. To a backdrop of the Rolling Stones and The Who, he demonstrated  that plants can use a form of RNAi to disrupt an introduced GFP gene, and that endogenous genes had no transactivity, compared to active transgenes, effectively discriminating between ‘self’ and ‘non-self’ genes

Researchers from the International Rice Research Institute (IRRI) are trying to improve rice breeds throughout Asia by selecting for desirable traits by QTL fine mapping, gene discovery, marker development, and molecular breeding.  IRRI researchers have successfully transferred the Sub1 gene into rice varieties aiding recovery from being submerged in flood waters for two weeks. Check out a cool time-lapse video demonstrating this, here. Recently, genotyping by sequencing (GBS) for GWAS on rice bacterial blight has been implemented. IRRI utilize a trait analysis by association, evolution and linkage (TASSEL) GBS ‘discovery pipeline’ and have developed user-friendly SNP data formatting and analysis tools through the online Galaxy bioinformatics resource. IRRI are working to refine the databases and informatics tools for analysis of NGS and high-density market sets to enable a new era of precision rice breeding.

Moving Forward

On Day 3, Illumina Marketing and development representatives presented their vision for moving forward in clinical research, agricultural and applied markets as well. This meeting has been a great opportunity for Illumina users to grasp the scientific scope of projects underway, and to bring corporate visibility to the region. The caliber of the talks was amazing, and clearly explains why Asia-Pacific countries represent some the fastest growing adopters of new sequencing and array technology– a real sign of how our region is coming into its own in this area of science. We trust you have found this account of the meeting of interest and we hope to see you next year!

 Image credit: Michelle Garred


Comment

  1.