Blog @ Illumina
Real scientists. Real commentary.

ESHG Wrap-up: Complex Disease, Rapid Diagnosis, and Personalised Medicine

Scott Brouilette, Ph.D.
| Jun 11, 2014

ESHG 2014 wrap-upThis last ESHG post will cover a few topics, not necessarily in chronological order, but research talks that struck me as being particularly interesting (and ones that we haven't covered in earlier posts). There was certainly no shortage of these...

Genetics of Complex Disease 

Eleftheria Zeggini from the UK spoke on the study of anthropometric traits, focusing on body shape and composition, using WGS and samples from both ALSPAC and TwinsUK. A large number of novel loci were identified, including a Chr5 waist/BMI signal, fine-mapping of a Chr11 height signal was achieved, and confirmation was provided for many existing loci.

As we all know, there is a wealth of GWAS data out there, much of which failed to identify the number/quality of hits expected. To date, more than 130 loci have been associated with cardiac electrophysiology, but as phenotypes become more granular, the sample size necessarily decreases. Jessica van Setten has taken a sub-set of available GWAS data and used larger/denser imputation reference panels to identify new loci associated with ECG traits. 52 genome-wide significant locus-trait associations were reported from 30,000 subjects, including 6 novel associations. Replication was performed in 22,000 subjects from the cohorts for heart and aging research in genomic epidemiology (CHARGE) consortium imputed with 1000 Genomes data. Jessica reiterated that new SNPs were identified due to higher density, better coverage of low frequency variants, and better haplotype quality (phasing). Overall, it is a great example illustrating the complimentary nature of SNP arrays and next-gen sequencing.

Autoimmune and chronic inflammatory complex traits clearly have genetic components, but the results of GWAS have been limited. Whole-genome sequencing may be the desired method to study these, but it is currently too expensive for many labs. Andréanne Morin introduced ImmunoSeq, which encompasses targeted sequencing of regulatory + exome + HLA regions to give a total target region of 138 Mb. Using this approach, they identified an average of 67 novel “regulome” and 69 exome variants that had been missed by the 1000 Genomes Project. Furthermore, they determined that their rare novel variants have high potential for functional impact via TF-binding site disruption, are globally enriched for enhancer RNAs, and increase allelic variance in gene expression. Future plans include an update to ImmunoSeq panels and a study of the impact of rare variants on chromatin states.

Implementation of NGS in Diagnostics

Pascal Joset started by stating that whole-exome sequencing (WES) has a diagnostic yield of 25-31% and asking if WES should then be used for all patients, or if panels are better in certain situations? He highlighted potential drawbacks of WES, such as lack of exon coverage, but showed that diagnostic yield could be as high as 59% when a clinical phenotype was suspected. On discussing the utility of the TruSight One Panel with 96% target exon coverage, Pascal showed that his diagnostic yield increased to 71% with better coverage than WES. So for cases where a clinical phenotype is suspected  TruSight One is used; in undiagnosed cases WES is implemented.

The dilemma over WES or WGS was touched on by Gils van Santen in the context of intellectual disability (ID); are there potentially variants that are missed by exome sequencing that are picked up by WGS? The instinctive answer may be yes, but Gils showed that, at least in their comparison, these are predominantly false positives and that “nothing special” appeared to be missed by WES versus WGS. In terms of output statistics, 140x coverage for the exome was “as good as it gets”, with 50-60 million reads giving >95% of the exome.

Gert Matthijs from EuroGenTest, one of the organizers of the satellite meeting discussed previously, outlined their guidelines for diagnostic use of NGS. These are largely based on the points raised, and the formal report will be available on the EuroGenTest website in the next few weeks for comments first, and subsequent publication by the ESHG.

After a stunning talk from Stephen Kingsmore on his critical work using sequencing technology for rapid analysis of neonatal inherited diseases at Children's Mercy Hospital, we then moved to rapid DNA sequencing in public health microbiology. Claudio Köser from Sharon Peacock’s lab asked: when does a technology become truly disruptive? When it is superior, faster, cheaper or simpler? Claudio believes that the last, simpler, offering economies of scale, is the most important for microbiology. Four key stages of testing where WGS can be applied include pathogen detection, pathogen identification, susceptibility testing, and epidemiological analysis. These testing phases represent a continuum, with the vast majority of samples coming up negative, but for positive samples the turnaround time and cost increases significantly as they move though testing pipeline. Claudio then described a single-colony WGS pipeline on the MiSeq System using Nextera library prep that returned results in less than 25 hours. While single-colony WGS is not currently cost-effective at the identification stage, he then revisited a now classic example of tracking a MRSA outbreak that lead to the closure of a neonatal ward.

Towards Genomic Personalised Medicine 

Paul Pharaoh took us “beyond BRCA1 and BRCA2”, referencing the collaborative oncological gene-environment study (COGS) on the custom Illumina array. COGS increased the number of loci for breast cancer to 72, explaining around 14% of inherited risk, and showing that alleles interact multiplicatively, resulting in a large number of genotype combinations that may confer risk. Genes included BRCA1/2, but also many others (such as PALB2 and ATM), and to date approximately 1/3 of the genetic component of risk for breast cancer has been explained.

Age-related macular degeneration (AMD) was the next subject, one that is always of interest to me, as my DTC genotyping test from two years ago indicated that I have a relative risk of 55% (versus 5% in the general population!) Caroline Klaver reminded us that AMD has both genetic and environmental influences, and told us about WGS analysis in families where inheritance could not be explained by known alleles, introducing rare alleles that can lead to early-onset AMD. She then highlighted the commercial tests that are currently available, but only 4/5 of the 26 known genes are routinely tested for, and her own risk varied considerably between 1.4–16%. So is genetic testing for AMD clinically relevant? The answer at the moment may be “no”, with the standard advice being to modify your risk by not smoking and eating foods rich in antioxidants. Solid advice.

All in all, ESHG was a fantastic and hugely diverse conference with top-notch scientific sessions that always spark new questions and lines of investigation. Thanks for reading these posts, and I hope to see you at a future genomics conference!