When even well-posed questions are hard to answer

On Friday, in acknowledgement of Rare Disease Day, our daughter Ellen blogged about living with a rare disease.  She wrote eloquently about her wish to understand why she has this disease, including, if it's a single gene disorder, knowing the causal variant.  She wrote about the advantages of this when navigating a medical system that isn't always sensitive to rare diseases, but in which genetics has become the gold standard.  We fully support her wish to understand why she has this disease, and have tried to help as much as we can. We would do the DNA work ourselves if we could.

Even so, she mentioned that her parents, Ken and I, are skeptics about a lot of genetic research.  Yes, that's true, but another word for that is 'realist'.  We are alive at a time in history when more is known about genes and genomes than ever before, and for decades we've been hearing promises of what this new knowledge will mean for medicine, and the promises roll on.  Once we all have our genomes on a disk, we'll be able to predict and treat whatever it is our DNA foretells.

Lazuli Bunting; rare birds in Central Pennsylvania; Wikipedia, Leander Sylvester Keyser

Except, except, Ellen's genome is on a disk.  Or at least her exome, the protein-coding parts of her genome.  Her disease, hypokalemic periodic paralysis, is one of several forms of periodic paralysis, which have been found to be associated with three different ion channel genes.  At one time a researcher in Germany was offering free genotyping to anyone diagnosed with the disease.  Ellen sent  blood samples, but was told that she doesn't have any of the known causal variants in these genes. She was also involved in a large whole exome study of unexplained Mendelian disease, but all they were able to tell her was that she doesn't have any potentially causal de novo mutations, mutations that neither Ken nor I have.  She is the only family member with HKPP, and as such, the initial question in a search for the cause is whether she has a variant that we don't have, that might be responsible.

And, to her frustration, that is all she knows.  It has been suggested that she go the clinical genetics route, having her DNA tested for known causes of HKPP, but that seems unlikely to be helpful, given that she knows what disease she has, just doesn't know why, and clinical labs don't look for new causal genes or variants, but instead a battery of those that are known.

Ellen has classic symptoms and classic triggers, and her disease is pretty well controlled at the moment, so identifying the cause, as she wrote in her post, might not change her treatment, but it would ease her mind about future dealings with the medical system.  As importantly, it might help future patients avoid the lengthy, destructive diagnostic odyssey she herself experienced, which itself would be a very satisfying outcome.

Big Data advocates will say that the problem is that not enough people with HKPP have been sequenced, and once we've got a million genomes or more, that will facilitate identifying Ellen's and others' causal variants.  But only 1 in 200,000 people have HKPP, so one million is unlikely to help.  And, though the data are rather sparse, some estimates based on those data suggest that a fairly large minority, a third or so, won't have one of the known causal genetic variants.  As with most diseases, the phenotypes vary greatly, and again as with most diseases, this is likely to be because every genome is unique, and genetic background matters, along with exposure to other triggering factors.

Perhaps there's an as-yet unidentified gene that would explain many of the unidentified cases, or there are many unique pathways to the disease, or both, but given the rarity and the heterogeneity of the periodic paralyses, it would take a huge amount of luck for even a large database to answer Ellen's question.  We should perhaps call it dumb luck, because the investigators vacuum up generic data without specific regard to, say, the physiology of this particular disorder (and the same for countless other disorders).  Of course, collecting data on every possible physiological or environmental factor, mostly with weak individual effects, isn't possible and that is a dilemma for modern public health science.

In addition, it's known from affected families that penetrance of alleles related to the periodic paralyses is not 100% -- some people with a 'causal' variant never experience an attack, making associating genotype with phenotype even harder.  Again, genetic background may affect this but, as with many genetic disorders with variable penetrance, it's not at all clear.  Incomplete penetrance is a fact, but also a fudge factor, because it leaves the impression the trait really is 'genetic'; in fact, we often don't know how many people have such mutations but no symptoms at all, because they aren't screened (but some studies looking for such asymptomatic cases have easily found them, and they can be as common as the 'causal' mutations in affected patients).

Further, it's possible that there are non-ion channel related causes of these channelopathies.  That is, something upstream is going wrong.  In that case, it's unclear where to even begin to look for genetic causation.  Thus, hypothetically in this instance, ion channels respond to the ionic concentrations inside the cell and in its environs.  Factors that affect the ion concentrations themselves could lead to effects similar to ion channel defects per se.  Thus, again just surmising, there are known environmental stimuli for attacks but these may affect the ion concentrations themselves, not the channel protein function.  And, of course, both could be at work, which would be rather expected given the many precedents for disease complexity.

And, it's possible that Ellen's disease is polygenic, or not genetic at all, though given that many cases of periodic paralysis, including in families, seem to have a single genetic cause, this seems unlikely.

Genetics asks two basic questions: What causes disease X?  And, who will get it?  The promises of the past few decades are that answers to both these questions are just around the corner for most diseases.  The NIH Office of Rare Disease Research reports that there are 7000 known rare diseases (diseases that affect fewer than 1 in 200,000 people).  The cause of many of these diseases has been identified, and by some criteria over 6000 specific genes have been associated with some usually rare single-gene disorder.  In many cases, it's possible to predict who will get the disease, and that is where genetic counseling is so useful.  It is, in our view, also where our limited research resources should be directed.  

But, if you read MT at all regularly, you know what we think about the promise of predicting common, complex diseases with genes.  Current science is very far from answering the two simple questions, what causes common, complex disease X?, and who will get it?  And, you know that we think that's because these questions can't be answered in any way approximating the promise of, say, precision medicine.  

But single-gene disorders are a different kind of problem.  What causes Ellen's HKPP? That seems to be a well-posed question, and should be answerable.  But to date, it hasn't been.  Labs are reporting 25-30% success with identifying the cause of rare genetic diseases (some somewhat higher success rates), so she is not at all unique.   We commented last week on the problem of identifying specific at-risk subgroups more effectively than blanket epidemiological studies currently can.

Are we skeptics?  Or are we realists?  When even the 'easy' cases, like Ellen's, the low-hanging fruit, are hard, what does this mean about the promises for genomics?  

Mendelian Inheritance: Basic Genetics or Basic Mistake? Part III.

Mendel

In his experiments, published in 1866, Gregor Mendel crossed two strains of domestic peas.  'Crossing' means that one parent was from each strain, and because the strains were inbred, this meant that there was little variation in the genomic backgrounds within each strain.  Mendel used different pairs of strains for the different traits he studied.  Further, he found that his chosen traits (seven of them, like round vs wrinkled or green vs yellow exhibited what he called 'dominance' (as we translate the term today).  That means that the effect of the allele (genetic variant) from one of the two strains was always manifest as its corresponding trait.  To revisit what we said earlier in this series, in the first generation of a cross, every parental pea plant was either GG or YY genotype (green or yellow) at a given test gene so every offspring plant inherited a G allele from one parent and a Y from the other, meaning they had the GY genotype and the peas and pods were yellow (in which case we say Yellow is dominant over Green).

But things are more complicated in the generation produced by crossing these plants, because then a specified fraction of each offspring type would be expected (as we noted in Part I, for GY x GY, the famous 1GG, 2GY, 1GG ratios were expected.  Mendel went a few generations beyond that,  and the ratios became more subtle, but the point is the same.  However, each generation allowed some scrambling of the genomic background of the strains, that is, at the rest of the genome, due to what is known as 'recombination' among the two parental strains' chromosomes.

It was in the context of those backgrounds with their limited variation, that the plants seem to breed 'true'.  Of course, there was statistical variation in the frequency of the relative offspring types (and Mendel was accused of fudging some figures to make his story come out closer to what he expected).  He did not get exactly the expected proportions.   But this variation, which is observed in every 'Mendelian' situation in any species, has always been attributed solely to chance allele transmission from parent to offspring (or to data fudging by tossing plants that didn't fit).  The fudging accusation is highly debated (see my paper, Goings on in Mendel's Garden, Evol. Anthropol., 11:40-44, 2002). But there may be a more serious issue, hidden in the statistics.

That issue is the assumption, from Mendel's day to today, of the expected proportions of plant types (green vs yellow, smooth vs wrinkled).  That expectation was based on pure, 100% dominance, that is on the inherently dominant physiological effect of the two alleles in any of one of his test crosses.  The assumption is that there is no genomic background variation that causes deviation from these 'pure' expected proportions.  We don't know to what extent there were 'greenish' or 'yellowish' peas, or peas with nondescript or mottled nature in Mendel's experiments, that would have been tossed out on the grounds of foreign pollination or whatever.  We know however that the variation was small enough that the assumption of dominance was good enough--for Mendel's purposes. 

At least one of his traits, plant height, was largely quantitative and less clear to judge (this was written up in the early 1900s, especially by OE White in a series of thorough papers).  We know that genes in the wild have many different variant states, not just two. And these have variable effects.  And this is true for Mendel's traits.  This is the same story, consistently, with variation at alleles related to human diseases (like cystic fibrosis, or PKU, etc.), and variation and genetic control in essentially any species carefully studied.  For these reasons, Mendel probably could not have done what he did with random samples of wild plants--anymore than we can do it with GWAS and complex diseases today.  Indeed, we must say that the above-cited paper raised all of these points, before the mountain of confirming data that subsequent studies generated and that we have today was available.  Of course, these are inconvenient facts if you hunger, naturally perhaps, for simple answers to fond dreams of perfect crops, and immortality through genetics.

In that sense, Mendelian traits are an artifact or illusion of his simple experimental set up, one he intentionally chose because the traits 'worked' the way he wanted them to. But there are further reasons than natural variation in the test genes themselves, for thinking that Mendelian inheritance has been, from the beginning, a very misleading notion.

This is the fundamentally mistaken notion that a gene is the same as the trait it contributes to.  It is the assumption of causal inherency.  Instead, what we know very, very clearly is that with a few kinds of exceptions (such as many lethal dysfunctional mutations in genes), the effect of an allele is contextual:  it depends on the environment and, in this case more importantly, on the genomic background.  That is, the variants at the many other genes in the same plant or animal affect how the allele of interest is manifest.  This is because no gene is an island, despite our clinging to Mendelian concepts of inherent causation for the last 150 years.

The degree of this contextual dependence varies from gene to gene, trait to trait, population to population, and species to species.  There is no single biological theory (other than, perhaps, this generalization) that predicts what we will find.  Some alleles in some situations act in a way that would make Mendel smile.  But few traits are, overall, like that.  10% or so of known devastating mutations in humans, that typically are called 'Mendelian', are the normal allele in other species!

To a great extent Mendelian inheritance of traits, that has become the sacred icon of modern genetics, is simply wrong!  Certainly, in any species one can find traits that segregate in the expected fashion to a degree satisfactory for the purposes at hand.  There is a spectrum of effects, and some are of this simple-enough causation.  But from a point of view of an actual theory of biology, it is the specturm not its extreme, that is important.

In this sense, genetic effects are as relative to each other as motion of objects is to Einstein.  This means that a subtle but centrally important point,  that there is really little if any difference between 'physiological' and 'statistical' dominance!  These terms were introduced earlier in this series of posts.  The idea of inherent dominance has been a misleading oversimplification from the beginning.  Dominance is just a sometimes-observed approximate correlation between alleles and traits that applies to a particular population.  Inherent biological dominance is an experimental illusion.

Traits in Nature, like diseases, that appear to be Mendelian, are those that in current circumstances, chosen among countless traits that have been studied, seem to appear in families in the classical way.  We know this very well, but it's not in most peoples' perceived self-interest to face up to it.  Even the more devastating mutations in 'Mendelian' single-gene diseases typically have variable effects.  The same mutations in mice are very often strain-specific.  And many alleles at the gene have less, and even more, variable effects.

For example, most individuals with 'recessive' diseases are not homozygotes for 'the bad' allele: they are heterozygotes for various alleles that compromise the trait to various degrees away from 'normalcy'.  
To account for this, but still to cling to our Mendelian paradigm, we introduce fudge factors to account for incomplete dominance and the like, that must be introduced in genetic (family) counseling risk estimates.

To the extent these statements are true, the idea of Mendelian inheritance (of traits) has been a stunningly misperceived, mistaken theory that continues to cost huge amounts of money for chasing genes 'for' particular traits, as if such genes have inherent causal, and hence inherent predictive properties.

Fear for sale? Are DTC genetic testing companies selling anything else?

Direct-to-Consumer Genetic Testing
Two opinion pieces appear in this week's Nature (here and here) on the problems facing the direct-to-consumer genetic testing industry.  These are companies, such as 23andMe, deCodeMe, or DNA Direct, that sell estimates of the likelihood of a consumer having a genetic disease or trait.  Obviously, what is on offer, and what people mainly want to know about, is what to be afraid they might get.  Fear for sale is not too far off the mark for this industry.

The industry has been growing fast, as more and more genes 'for' traits and diseases are published, but legal and scientific questions about what these companies have to offer abound.  

The burgeoning, but virtually unregulated, direct-to-consumer (DTC) genetic-testing industry faces some serious changes in the United States. In a series of hearings last month, the US Food and Drug Administration (FDA) hinted that it will impose new regulations on companies selling such tests. The agency has also sent letters to test makers, as well as to one maker of the gene chips on which many such tests rely, saying that the firms are not in compliance with its rules.

In addition, the Government Accountability Office (GAO) last month unveiled the findings of its year-long investigation into the scientific validity, safety and utility of the gene tests used by the industry. The report called some of the tests misleading, pointing out inconsistencies in the results they provided, as well as some companies' shady marketing practices.

Not all companies have been found to have shady marketing practices, or to intentionally mislead, but many of the other issues pertain to all of these companies because they pertain to both the science and the product they sell.  Indeed, it can be argued that the probabilistic nature of these tests means that they all mislead, even if unintentionally, especially as so few people have a good sense of what probability means.

Regulation

Genetic-testing services are proliferating fast. In 1993, tests were available for about 100 diseases. By 2009, the number was almost 1,900. Some forms of testing are major advances in the diagnosis of certain conditions, such as Rett syndrome and types of brittle bone disease. The clinical utility of others — such as the high-throughput genotyping that is widely offered by companies that sell tests directly to consumers — is debatable.

The problems with these companies are multiple, some having to do with the marketing of tests, and some with the quality of the testing and the test results themselves.

According to the FDA, these companies are performing medical tests and marketing medical devices without a license, and this needs to be corrected.  There's no doubt that's how many people view these tests, and how they interpret the results.  As such, the FDA believes it should regulate these companies as it regulates any company selling medical devices or medical tests.  The FDA has recently brought this to the attention of many of these companies -- here is the letter sent to 23andMe on this subject in June, e.g..   

But, many observers believe that the way the testing is done also needs to be regulated to standardize DNA sequencing results, for example, and otherwise assure that results are valid.  So, that's also something that's under consideration.  According to the Nature piece, less than 1% of DTC genetic testing is regulated in the UK, and the proportion is low in the US as well.

Science 
While regulatory and legal issues are real and must be thrashed out, a deeper and thornier issue has to do with the science -- the actual causal connections (if and where they exist) between DNA sequence differences and a particular disease or other trait, and the assumption that we have accurate knowledge of those connections, which is usually far from the case.

If you send your DNA to 3 of these companies, you'll get not only risk estimates for 3 different sets of diseases and traits, but estimates you're given for the same diseases can differ.  Yet isn't there just one truth out there?  The  problem arises because the study results these companies are basing their estimates on can vary considerably, depending on who's included in the study, how the trait is defined and so forth, and which results the company chooses to use to calculate their own estimates will determine the estimated risks they send to you.  Further, there are often in fact no correct answers when it comes to complex diseases -- your risk of type 2 diabetes, or heart disease or stroke is genetic and environmental, and your particular genetic component is unique to you, and exposure patterns are always changing, which all means it can't be necessarily be accurately predicted from a pool of other people's genes.

Different genes and different alleles contribute to risk in different populations, and in every individual.  And, risk of complex diseases is due to the contribution of more genes than have yet been identified, or even that will ever be identified for you.  Even if all the genes that contribute to risk of type 2 diabetes could be identified for your neighbor, or even your parents!, the list won't be the same for you.

And this is before we even begin to consider the environmental contribution to your risk of disease, and this can be very elusive.  One example is risk of breast cancer in women with BRCA1 or 2 mutations -- mutations which confer some of the highest risk of cancer known.  Risk is very different depending on whether a woman was born before 1940 or after, because of changes in exposure to environmental risk factors.  And, we can't predict future environmental risk factors, so it's impossible to know what gene by environment risk will be going forward, which is what these companies in effect are doing.  Yet recent history very, very clearly shows that secular trends in risk, which must be due to lifestyle exposure differences, frequently make huge differences in risk.

The solution?
Most results of DTC genotype testing amount in some way to fear for sale.  That may not be the companies' intent, but it's not unfair to describe them in this way, because it's increasingly unlikely, as more data accumulate, that anyone will be given a no-worry clean bill of genetic health.  Everyone is at risk for something!  Especially if the criteria for expressing risk are not particularly major or definitive.

And, as we've said, the risks they are selling are mainly pretty elusive.  But when it comes to real and substantial risk, we've already got a system, called genetic counseling, for estimating risk for hundreds of known single gene disorders.  These are genes for which risk and predictability are extremely high, and that are primarily pediatric disorders like Tay Sachs or cystic fibrosis, though Huntington's, which strikes in adulthood, is another example.  Prediction of these disorders has long been done by genetic counselors, who are trained to predict risk (which, with these disorders, is fairly simple to do, as they generally follow Mendelian rules of inheritance) and to inform and counsel people about their risk.  They work in closely integrated ways with clinics, physicians, and medical schools.  Counselors are professionals who are tested, regulated, and must keep up their license with regular education.  So, where risk is actually definable, we've got no need for DTC testing. 

If genetics was still in the business of finding genes that follow Mendelian rules, they'd just be added to the list of traits that genetic counselors understand and counsel about.  But, largely driven by the profit motive it must be said, the genes these days being identified 'for' complex diseases by and large have individually small effects, effect estimates vary considerably by study, and they are simply not now, or probably never will be useful for accurately predicting your risk of complex disease.

So, how the DTC genetic testing industry does its testing is one question, and the FDA is trying to decide how to regulate this.  Whether the industry has anything of real value to sell you is another question entirely.  That's the rub.