Maps for the world of genomic medicine: The 2011 CSHL Personal Genomes meeting
The fourth Personal Genomes meeting was held at Cold Spring Harbor Laboratory, New York, from 30 September to 2 October and provided an exciting collection of science built on recent significant milestones in individual human genome sequencing, from the first personal genomes to thousands of human genomes sequenced. As ultra-high throughput sequencing platforms enable the production of more and more individual genomes, a growing number of scientists, physicians, and clinical geneticists are actively exploring the promise and the implications of these new data. Personal Genomes brought many of these pioneers together with nearly 200 scientists, physicians, ethicists, and others to discuss the progress and opportunities around the sequencing and medical interpretation of individual genome sequences. Hum Mutat 33:1016–1019, 2012. © 2012 Wiley Periodicals, Inc.
Personal Genomes 2011
Cold Spring Harbor Laboratory in New York was the site of the fourth “Personal Genomes” meeting. The meeting, held from 30 September to 2 October 2011, was organized into five main scientific sessions: personal cancer genomics, medically actionable genomics, personal genomes, rare diseases, and clinical implementation of personal genomics. A separate 3-hr session was dedicated to a panel discussion of ethical issues relating to return of results and there were two dedicated poster sessions in which a wide variety of relevant work was presented.
While the topics were wide ranging across the sessions, two main themes emerged. The first theme was the use of and the potential for individual genome sequences and next generation sequencing (NGS) technology to facilitate understanding of the genetic basis for phenotypes. The second theme related to the practical considerations of clinical genomic sequencing from the progress in deploying this technology for patient care to the issues associated with return of results to patients and other sequenced individuals. This meeting report is organized around these common themes and will draw on presentations from all of the sessions to illustrate how Personal Genomes 2011 helped to provide the basic maps for the coming world of genomic medicine.
Theme 1: Individual Genomes and the Discovery of the Genetic Basis for Phenotype
With the drastic increase in throughput and decrease in cost, NGS has reshaped the research landscape of molecular genetics and increased the speed of the discovery of genetic mechanisms for various phenotypes, especially disease traits. Many talks throughout the conference reported exciting discoveries made from the availability of individual genome sequences. Here we highlight some.
The opening keynote presentation by Dr. Elaine Mardis (St. Louis, USA) gave a thorough overview of the progress of NGS-based individual genome sequencing in cancer genome research in the five short years from 2007 to 2011. In addition to discovery and validation of many tumor mutations in different cancer types, deep sampling of genomic variants generated by NGS allows deep digital sequencing analysis, a statistical method to study the heterogeneity of tumor genomes present in a single sample. The analysis identifies the number of tumor subpopulations or subclones and distinguishes the clonality and allele frequency differences between primary tumor and relapsed tumor, permitting a model of tumor genome evolution with profound implications. Using real life examples, Dr. Mardis continued to demonstrate the power of clinical whole genome sequencing (WGS) in disease gene discovery. In a therapy-related acute myeloid leukemia (t-AML) patient, WGS identified a novel TP53 germ line deletion that was heterozygous in skin samples but homozygous in tumor and explained the cancer susceptibility of the patient [Link et al., 2011]. In another example, an AML patient displayed pathology characteristics of acute promyelocytic leukemia (APL) and had a complicated cytogenetics presentation, but not a cytogenetically obvious pathogenic PML–RARA fusion. WGS revealed a cryptic fusion of oncogenes PML and RARA due to an insertion event, rather than translocation. This discovery permitted the patient to be treated for APL rather than having a stem cell transplant, thereby avoiding a high cost procedure with high morbidity and mortality risk. So far the patient's outcome is promising [Welch et al., 2011].
Dr. Karin Kassahn (Brisbane, Australia) provided a natural follow-up to Dr. Mardis' keynote in the area of pancreatic cancer. Extensive genomic studies of one pancreatic cancer patient used WGS, whole exome sequencing (WES), mRNA and miRNA sequencing, and microarray analysis of methylation and copy number changes to identify somatic mutations and aberrations. These included substitutions, copy number variations (CNV), loss-of-heterozygosity (LOH), and amplifications that have implications in the onset of pancreatic cancer [Jones et al., 2008, 2009]. On this basis, Dr. Kassahn and colleagues identified several key driver mutations that confer chemosensitivity to mitomycin-C, suggesting this as an alternative therapeutic reagent to the standard treatment. Tests of the effectiveness of mitomycin-C treatment for this cancer in matched xenographs were positive while standard treatment showed no effect.
Dr. Lisa Trevino (Houston, USA) described a cancer genome sequencing project consisting of hundreds of samples in seven different cancer types. The project will test the two-hit model of cancer susceptibility and hopes to identify mutations that may explain the cancer susceptibility. The two-hit model hypothesizes that in addition to a primary germ line mutation, a second somatic LOH mutation is required for tumorigenesis [reviewed by Berger et al., 2011]. From the extensive genomic sequence data they collected, Dr. Trevino and colleagues first identified loss-of-function (LOF) germ line mutations. Then from tumor genomes, they identified inactivating somatic mutations; some of them in genes that already carry a germ line primary mutation. Using this strategy, they were able to identify LOH germ line and somatic mutations in cancer suppressor genes such as CDH1, APC, ATR, and BRCA2 in different cancer types, suggesting the somatic mutation is the second hit that caused the cancer phenotype.
Structural variation in individual human genomes has profound medical implications. Dr. Kathleen Burns (Baltimore, USA) presented an intriguing study about human structural variations introduced by retrotransposon insertions. Dr. Burns described a method called “transposon insertion profiling” or TIP developed in her lab that can be used for detecting recent insertion events. The technique is based on ligation-mediated PCR amplification, taking advantage of signature sequences found in L1(Ta) LINEs and a subset of Alu SINEs. Using methods combining TIP with NGS or tiling microarrays (TIP-seq and TIP-chip), Dr. Burns and her colleagues were able to identify large number of insertion polymorphisms and add to evidence that retrotransposons are ongoing sources for genomic changes [Huang et al., 2010]. Interestingly, it was discovered recently by the Devine laboratory that L1(Ta) can act as somatic mutagen in cancer [Iskow et al., 2010]. Dr. Burns' group is now studying the impact of retrotransposon insertional polymorphisms (RIPs) on human phenotypes. They have found that RIPs commonly occur within short distance of trait-associated SNPs, suggesting RIPs in intergenic and intronic regions have inherent potential for functional effects.
Dr. Joris Veltman (Nijmegen, Netherlands) described a successful effort using WES to identify the genetic basis for Mendelian diseases. By carefully choosing well-phenotyped patients and families, Dr. Veltman and colleagues were able to identify causal mutations in dominant sporadic diseases such as Schinzel–Giedion syndrome [Hoischen et al., 2010] and Bohring–Opitz syndrome [Hoischen et al., 2011]. WES of patients with neurodevelopmental disorders such as intellectual disability (ID), autism, and schizophrenia lead to significant understanding of these disorders. For example, by exome sequencing of 10 patient trios with ID, they uncovered in each patient between zero and two de novo, validated mutations in nine genes, of which two are known to be related to mental retardation, and four are involved in neural system development [Vissers et al., 2010]. A larger study indicated that 40–50% of severe ID cases could be explained by de novo mutations and 10% by de novo copy number changes. Dr. Veltman highlighted their ongoing efforts to implement clinical exome sequencing to diagnose patients with genetic disorders.
Appearance is the most obvious human phenotypic trait and the ability to predict traits such as eye and hair color has important applications in forensic studies. Prof. Manfred Kayser (Rotterdam, Netherlands) presented results of a study where numerous SNPs previously associated with categorical eye color were tested for their predictive effects on eye color in thousands of Dutch Europeans. They found that just six SNPs from six genes led to greater than 90% predictive accuracy for blue and brown eye color [Liu et al., 2009]. As eye color truly is a continuous trait, Prof. Kayser and colleagues also carried out a genome wide association study (GWAS) on quantitative eye color extracted from digital eye images and found three new eye color genes [Liu et al., 2010]. In a prediction study of hair color, testing 46 hair color-associated SNPs from 13 genes in 400 Polish samples, Prof. Kayser and colleagues demonstrated that 25 SNPs are sufficient to successfully predict different hair color categories from DNA [Branicki et al., 2011]. Despite these successes, other physical traits such as human body height are much more complicated genetically and can therefore not yet be predicted from DNA with practically useful accuracies. For example, 180 SNPs uncovered by GWAS can explain only 10% of body height variation in the more than 180,000 subjects tested [Lango Allen et al., 2010], while 80% of body height is assumed to be heritable.
As a popular method to study genetic basis for phenotypes, GWAS have been widely used in the past decade. However, the alleles discovered by GWAS generally explain only a very small fraction of the genetic variance causing common diseases; the GWAS results cannot explain the large percentages of samples that do not have the associated risk alleles but still develop the disease phenotype. In his keynote speech, Dr. Richard Gibbs (Houston, USA) went deep into the complex architecture of human diseases and offered some insightful views. While noting that the “common disease/common variant” theory was the intellectual foundation for GWAS, he stated that the large number of whole genomes that have been sequenced demonstrate that the number of rare and private variants discovered is much greater than previously predicted. Importantly, evidence is emerging to support the idea that recent, rare mutations may have a greater influence in disease risk than ancient variants inherited from distant ancestors and that specific loci that are causal for Mendelian disease may contribute to the overall risk to specific common diseases; likewise, loci discovered by GWAS for complex traits may have known causal alleles for Mendelian disorders. One example is that, of 891 annotated genes with significant association in NHGRI's GWAS catalogue, 268 were found to have mutations in rare single-gene disorders [Lupski et al., 2011]. Dr. Gibbs presented a unified model for human genetic disease, which proposes that all major categories of genetic diseases—Mendelian, common, genetic disorders, and chromosomal syndromes—are explained by a continuum in which the same genes are affected by different types of genetic variations and through different modes of inheritance.
Theme 2: Clinical Genome Sequencing and Return of Results-–The Future is Now
Since the publication of the first drafts of the human genome sequence in 2001, a major research focus has been to translate genomic research into everyday medical practice. Ten years later, a steady stream of relevant reports indicates that we are now consistently applying knowledge learned from sequencing of individual genomes into the decision-making process of medical care. Indeed, the analogies for sequencing vividly draw on existing tools for clinical diagnosis: “Clinical WGS is the pathologists' new microscope” (Mark Bonguski), “Clinical WGS is the new X-ray” (Richard Gibbs). Through the many talks about clinical sequencing of patient genomes, the world of individualized genomic medicine is now coming into focus with the accompanying ethical issues related to returning results to patients and other sequenced individuals. In this section, we first summarize several talks about individualized medicine and return of results of clinical WGS to patients and primary care givers. We then highlight additional presentations that focused on the ethical issues of returning results.
The Saturday morning keynote speech by Dr. David Valle (Baltimore, USA) featured an overview of the history and current trends in clinical genomic sequencing and individualized medicine and identified WGS as critical to understanding disease mechanisms and disease genes. Dr. Valle cited the 2,600 disease genes recorded in OMIM and the more than 1,900 diseases with specific molecular tests [Reviewed in Beaudet, 2010]. He was confident is his prediction that essentially all Mendelian diseases will be solved as part of the International Rare Disease Research Consortium (IRDiRC) that includes an NHGRI-funded effort at several U.S. institutions. Dr. Valle introduced Johns Hopkins' Individualized Medicine Program that acquires and banks sequence information (SNP, WGS) for individuals and carries out functional studies for disease linkage now or in the future. With these efforts, the journey toward the era of genome-informed individualized medicine is underway.
Dr. David Dimmock (Milwaukee, USA) talked about lessons learned from his experience of clinical sequencing to identify the molecular basis of diseases and modify disease management. In one case, a 4-day-old patient suffering from a severe genetic disorder at birth was found to carry two mutations in the TWINKLE gene and these mutations were known to be lethal in previous cases. With this knowledge, a planned liver transplant was canceled allowing the patient and the family to spend as much quality time together as possible before she passed away 6 months later in the comfort of her mother's arms. Dr. Dimmock introduced the clinical WGS program at Children's hospital of Wisconsin and the workflow developed by the program. So far, 47 clinical cases have been evaluated through the WGS program and it is expected that this number will grow considerably.
Dr. Wojciech Wiszniewski (Houston, USA) gave an example of a positive therapeutic outcome using information obtained from clinical WGS in the case of twins suffering from Dopa-responsive dystonia. Causal pathogenic variants were identified in the sepiapterin reductase (SPR) gene and the discovery suggested optimizing medical management with additional 5-hydrotryptophan (5-HTP). The L-DOPA plus 5-HTP therapies resulted in a further reduction of the neuropsychiatric phenotypes in the patients.
Returning genetic test results to patients and research participants requires the appreciation of a number of sensitivities. As introduced by Dr. Jennifer Ivanovich (St. Louis, USA) in the opening presentation of the annual Personal Genomes Ethics Panel discussion, communication of genomic information from genome sequencing to patients in a positive and proactive way is a subject of active research [e.g., Schalowitz and Miller, 2008]. When communicating the results to a patient, a number of considerations must be taken into account. These include where the discussion should take place; when to initiate the communication (i.e., while the participant is healthy or after they have become sick?); who should communicate with the patient: the primary physician, geneticist, or genetic counselor; what information should be provided: complete, selected, or targeted results; and how to communicate: a single counseling session or should follow-up sessions be considered?
Dr. Stephen Kingsmore (Kansas City, USA) and his colleagues at Children's Mercy Hospital have developed an NGS-based diagnostic test for 619 severe recessive childhood genetic disorders. Exon capture is used to enrich target DNA in samples to achieve a deep sequencing coverage for variant discovery and genotyping [Bell et al., 2011]. Such testing is being evaluated as an economical way to reduce the cost and time to diagnosis of severe recessive disorders as a starting point to address the suffering associated with these disorders; similar carrier screening tests for prospective parents are also being developed. In the discussion of the ethical issues around these tests, the scope for preconception carrier screening was suggested to be voluntary community-based population screening that has previously shown general acceptance from people without creating discrimination or stigma. For newborn and childhood genetic testing, return of results is a very sensitive task. Current federal guidelines stipulate that while all discovered variants should be reported to adults, carrier status or risk for adult-onset conditions is not to be disclosed to a child.
Dr. Mark Boguski (Boston, USA) described the nontechnical barriers for clinical genomics to become a routine part of medical care. The first barrier is the transfer of knowledge from research labs to patients. People who will be potentially involved in the workforce of clinical WGS diagnosis are medical genetists, genetic counselors, and pathologists. Together with clinicians, they will be responsible for communicating the lab results to patients and making diagnostic and treatment decisions. Genetic training of the existing workforce is a critical task that requires immediate attention. A stakeholder meeting in the field of clinical WGS was held this year in an attempt to outline a blueprint for genetics training including residency training in genomics for pathologists [Tonellato et al., 2011]. The official website for the initiative is http://genomicmedicineinitiative.org/. Other barriers include the legal and regulatory aspects of laboratory-developed tests (LDTs), insurance reimbursement, patient privacy and legal protection, and the lack of clinical-grade knowledge database. U.S. federal organizations such as the FDA and the Center for Medicare and Medicaid Services (CMS) will play major roles in developing and implementing appropriate rules and regulations to address these barriers.
The talk presented by Dr. Gholson Lyon (Salt Lake City, USA) included real-life struggles with whether and how research results should be returned to a participant. It is generally agreed that research results should be confirmed in a Clinical Laboratory Improvement Act (CLIA)-certified lab before being returned to a patient. However, the practical implementation of this is very difficult in the context of the new technologies of exon capture and sequencing. Dr. Lyon described two specific scenarios to illustrate this. In the first case, his group extensively studied families with a rare X-linked, infantile lethal disease, which they have named “Ogden Syndrome,” and identified a mutation that perfectly segregated only in affected individuals. During the course of the research, an expecting mother in one of the families was found in the research lab to be a carrier of the mutation. However, the process of developing a CLIA-certified test took many months, and Dr. Lyon and his colleagues could not communicate the research finding with the carrier, as the results were not obtained in a CLIA-certified lab. The baby was born with the disease and died from it within a few months. Since this episode, a CLIA-certified test for Ogden Syndrome based on Sanger sequencing has been developed at a major diagnostic lab, but the process of developing the test and then getting someone (family or insurance company) to agree to pay for the test has been quite difficult. To date, the family still have not received the test. The second case involved a family with attention deficit hyperactivity disorder (ADHD) in three men and idiopathic hemolytic anemia (IHA) in one of them. While Dr. Lyon and his collaborators have not yet proven causality for some of the rare genetic variants in the ADHD patients in the family, they discovered a causal mutation for IHA in the one family member. They conveyed the anemia result to the patient's haematologist, so that the doctor could validate the finding in a CLIA-certified laboratory, prior to any disclosure to the patient. However, once again, the implementation of this has been quite difficult, due to financial and other barriers and the man still has not been given any CLIA-certified test results. Dr. Lyon suggested that researchers strongly consider performing all exome and WGS up front in CLIA-certified laboratories, so that results can be more readily returned to patients and research participants.
In summary, advances in sequencing technology have led to the routine generation of whole genome and whole exome sequences in research facilities and the first steps into the world of individualized genomic medicine have already been taken. The analysis of sequence data continues to shed light on the understanding of the genetic mechanism of disease phenotypes; and more excitingly, it has started to generate results that have implications on disease diagnosis and treatment. Although the amount of remaining work is enormous, upon departing Cold Spring Harbor Laboratory at the end of Personal Genomes 2011, we are navigating with ever-improving maps of the world of genomic medicine.