We are all prisoners of our own genomes, but our genomes are cages that we cannot see. What if you were given a magic wand that could extract your genome from its shroud of invisibility? Would you choose to use it? And, if you did, would you be any less the prisoner?
Well, such a magic wand is now yours (in beta) for $1,000 or so, and these questions have become very real. The new reality of genomics as a tool for describing the individual, rather than a species, made for an inspirational set of talks at this year's Cold Spring Harbor "Personal Genomes & Medical Genomics" meeting (held on November 14-17), which saw a series of speakers echo one another in telling the story of a clinical genomics revolution. The past few months have been witness to a truly pivotal moment in the history of medicine: sufferers of mystery diseases, who for years had been poked and prodded in vain, are now being sequenced in their 100s and 1,000s, with a diagnostic success rate of between one-quarter and one-third of patients.
The "missing" diagnoses
Better news still is the optimism that this rate can be improved even further. The most common approach used in clinical genomics is exome sequencing (that's just the genes; the whole shebang is pricier and generally somewhat unfathomable to interpret), so it might be the case that some of the missing diagnoses are to be found outside of the exome. Such an eventuality is not too disheartening, because whole genome sequencing is predicted to become more affordable, and because, as a greater volume of data is generated, perhaps we will arrive at a more sophisticated understanding of the genome's dark matter.
More immediately, we can take comfort in the likelihood that a large proportion of the missing diagnoses actually lies within the exome data we already possess. A number of presenters strongly emphasized how difficult it is to determine which of the many variants to be found in a given exome contribute to a disease phenotype. While complementary sequencing of family members is frequently used to provide vital clues, alongside bioinformatics tools that predict loss-of-function and common variant databases, we still lack power in this area. The comprehensive cataloging of common variation and, crucially, a better understanding of gene function will be key to improving the confidence with which we can diagnose the genetic basis of disease from individual exome sequences.
Even at the current success rate, the power of clinical genomics is life changing to many human beings that had previously been viewed as undiagnosable: this is at heart a real human story more than a scientific one. In some instances, the discovery of a mutation was so informative that it led to a life-saving or life-enhancing treatment. The point was also made that a diagnosis in and of itself, without a consequent treatment, can be very valuable in providing peace of mind or closure to an individual, or reassurance of a genetic cause to families that experience misplaced guilt.
This principle can also apply to the use of genomics for understanding conditions that are already diagnosed: several presenters detailed work with autistic individuals that has collectively produced 100s of exomes and identified nearly as many putative autism risk genes. While the scientific goal of this work is to understand the biology of autism, which is hindered by the evident genetic heterogeneity at play, identifying "autism" mutations can be meaningful to the families concerned.
Quite the opposite is true when it comes to pharmacogenomics: this is a discipline that is very unlikely to alter your philosophical sense of self, but is of enornmous practical use, particularly in drug safety (but also in guiding the choice of drug for maximum efficacy).
Whose genome is it anyway?
While our ability to illuminate an individual's genome, and in some cases even liberate the prisoner trapped within, is undoubtedly now a force for good in the clinic, it also raises difficult questions and challenges that in many instances we are poorly equipped to answer. A common theme was the problem of so-called "incidental" findings – these are data that are not related to the disease or phenotype being studied, but that might otherwise be of interest to the individual. If a mutation is discovered that is known to confer a disease risk, should the patient be informed? How about family members, who might also have inherited the variant? What risk factor should the threshold for reporting incidental findings be set at? What about diseases for which no treatment or preventative measures are available? Would this information actually do more harm than good?
And who should decide the answers to these questions – the patient? The patient's physician? An ethical review panel? What if the patient is a minor – should the parents be party to genomic information, even if it will only affect the patient in adult life? What if the patient is a fetus – should a higher confidence threshold be required when genome sequencing is performed prenatally?
A related issue is the question of making personal genomic data openly available in repositories and databases. This will no doubt aid research, but does the openness of data represent an unwarranted invasion of privacy for the individual? How can we be satisfied that a patient's consent for such openness is "informed" when we are dealing with novel and complex concepts, and many unknowns?
Another key question is how to compromise between speed and accuracy when reaching a diagnosis is time critical. The accompanying challenge, of course, is to develop faster sequencing and analysis pipelines.
Personal sequencing for personal genomes
An elephant in the room at every genomics meeting these days is the promised, but as yet not materialized, Oxford Nanopore sequencer. While the new dawn of clinical genomics has been driven by immense technological strides in equipment that is already on the market, the hypothetical specs of the nanopore constitute the holy grail of sequencing. Indeed, the device is awaited by some with quasi-messianic fervor.
One exciting feature of the Oxford Nanopore project is the plan to make the technology available as a USB-sized "MinION" sequencer that can be run anywhere you can take a laptop or tablet device. When accompanied by the hoped for idiot-proof analysis software that many bioinformatics labs are currently dedicated to developing, the "MinION" would cut out the need for professional sequencing and analysis services, just as "Direct To Consumer" companies such as 23andme have already bypassed the need for clinicians in accessing your genomic data. So is personal sequencing of personal genomes the future? Will you, the prisoner, uncover the hidden depths of your genome in between levels of Angry Birds?