Another major gap from me as my travels have continued…
A major theme that emerges for me over the past few months is on non-medical uses of genetics, made possible by 1) Continued growth of Direct to Consumer (DTC) testing, 2) large scale studies of non-health traits, and 3) increased acceptance of polygenic risk scores, whereby an aggregate score for a given trait is made from small contributions from many genetic markers. This is something we need to talk about.
1) Millions of DTC tests have been sold in the US (over 12 million, according to one estimate from February). A poll found that “Some 17 percent of Americans already have undergone at least one kind of DNA test, and 52 percent of the remainder say they’d like to.”
2) In the last few months, there have been large scale studies of neuroticism, intelligence, social mobility, and on social traits including loneliness. A lot of research into such “social” traits is performed under a health mandate, as many such traits correlate with health outcomes.
3) Polygenic risk scores can be much better at identifying those at risk of serious conditions. That’s the conclusion of a paper and the basis for a new tool that will ingest e.g. 23andMe data and give you a score. From the paper: “The approach identifies 8.0, 6.1, 3.5, 3.2, and 1.5% of the population at greater than threefold increased risk for coronary artery disease, atrial fibrillation, type 2 diabetes, inflammatory bowel disease, and breast cancer, respectively.” These successes, combined with the type of studies I listed above, lead to, for example, educational achievement and cognitive performance polygenic risk scores, explaining 11-13% of variance, 7-10% of variance respectively. Such scores have already make it into DTC tests, for example educational achievement markers are available from Helix.
Needless to say, such scores have the potential for large social implications. The educational achievement study has appeared all over the press, but I have found good critique of possible implications scant. This piece is an exception, one example:
“There’s as much to be learned about the nature of our educational system as about the nature of the individual in this data,” said Mary Helen Immordino-Yang, professor of education, psychology, and neuroscience at the University of Southern California. Specifically, if certain genetic variants are associated with better educational outcomes, then there might be something about the structure of our educational system that’s favoring people with these variants. For example, if the variants were involved in language comprehension, that could tell educators that current teaching methods aren’t working for students who process language differently. That means they should be designing new interventions to accommodate that variation, Belsky said.”
Such reflection is critical. Contrast this with the authors of a paper titled “The stability of educational achievement across school years is largely explained by genetic factors”, state the motivations and consequences of their work are that “we could use DNA tests at birth to identify children at genetic risk for developing reading problems, and give them early intervention.”
Separately, additional potential downsides of the DTC genomics movement have been in the news. I highlight three:
- What happens when DTC tests reveal unexpected family secrets? The Atlantic has a piece on a facebook group, with over 1000 members, that allows individuals affected to vent emotions and plan next steps. Meanwhile in the UK, the fertility regulator has called for genetic testing companies to better highlight chances of uncovering family secrets, additionally saying that anonymity for sperm/egg donors is a thing of the past.
- In the light of Ancestry.com and Spotify teaming up to offer playlists based on your results, a critique of how genetic ancestry testing, who focus on the fact your DNA reveals something meaningful about you. This runs from “conflating DNA and cultural identity” (routing for a World Cup team based on results) to “game programs set up to address past injustices” (using results to prove Native American or African ancestry), to reifying race as a meaningful category (citing a study that showed reading about these tests increased beliefs in racial differences).
- A report on four cases of families who act on information obtained from raw data provided by DTC tests. In these cases, none of the SNPs reported turned out to be present. Another story of a discordant 23andMe and Ancestry result over a very worrying variant.
I think it is worth highlighting what we have learnt about consumer preferences
- The Associated Press reports a poll of 1109 adults it performed on questions related to genetic testing (also refed above on number of people interested in genetic testing). On whether people would want to know if DNA showed they had a genetic variant associated with an incurable disease. 60% said Yes, for under 30s the number was 78%. Most (but not all) would tell family members). The poll also asked about the use of DNA by the Feds — “Half of people think genetic data should be used to help solve crimes only with the consent of the person tested, a third think it’s OK without that consent — and 13 percent don’t think law enforcement should use it at all.”
- A Pew poll found majority support for gene editing for babies for health reasons. Men and more supportive; religious people are less supportive. If the intervention relies on testing embryos, most are opposed.
- A U of Michigan study on patient attitudes to their biobank samples being commercialized. “67 percent of participants agreed that clear notification of potential commercialization of biospecimens is warranted, but only 23 percent were comfortable with such use. Sixty-two percent believed that profits should be used only to support future research, and 41 percent supported sharing profits with the public.”
- STAT reports on consumer adoption of genetics in China. Those in the space refer to a particular emphasis on a Chinese fascination of how genetics affects identity and destiny, and, for the generations born under the one child policy, with finding family. Routine newborn genome sequencing is on the cards within the next five years (Veritas already has a product in pilot out there). Because there isn’t an entrenched medical genetics profession, there will be less paternalism about results.
It strikes me that the academic literature is all about efforts to ensure that people’s informed choices are respected. A lot of this type of work, which looks at what those choices are likely to be, happens outside the academic mainstream. This is an observation I intend to check.
Applications
- A Federal official said that DNA tests were being used to reunify families. Jeff Sessions says on this clip from Washington Watch that many of the children may not be biologically related to the adults that bring them to the border. Ethical issues abound.
- The UK’s Genomic Medicine Service will launch on October 1. All new cancer patients will have their tumors genetically sequenced.
- An interesting story about how the existence of an ICD code to cover a condition can have a direct impact on patients’ lives — these codes are what is searched in databases, and what a lot of insurance is tied to.
- 23andMe and GSK have entered into an exclusive drug development partnership, and GSK have bought a $300m stake. They hope to use 23andMe’s database to locate better drug targets. Note that at about 5m customers, that is ~$60 per current customer. In possibly related news, 23andMe have blocked access to deidentified raw genetic data by commercial third parties through its API.
- A firm is inviting you to upload your genetic info to help you plan for your financial future.
- Many DTC companies will adopt new privacy guidelines. These include companies getting explicit consent from customers before sharing with third parties. Though the title of this piece captures it: Deleting Your Online DNA Data Is Brutally Difficult.
- Australia has a government run, opt out electronic health record, which will be able to store genetic information. Some fear for privacy.
- The first CRISPR based gene therapy trial backed by a US company has started. It will recruit beta thalassemia patients in Germany.
- George Church has his say on how we benefit from having our genomes sequenced. He introduces a new start-up he has co-founded, Nebula, which will use blockchain technology to address “issues of transparency and control” in “delegating data access and recording of transactions”. He states: “The Nebula team and approach I personally trust because of emphasis on ethical issues.”
- A NYTimes Op-Ed that successes of precision medicine and vastly outnumbered by failures. At the ASCO conference, the four precision therapy trials presented with results were all failures, and yet there was no news reporting.
- Federal money to develop a system intended to recognize emerging bio threats based on DNA orders placed, to e.g. spot if someone was trying to re-create an extinct virus.
Regulation
- A Federal Court maintained the status quo over IP for CRISPR in eukaryotic cells, maintaining the patents of the Broad Institute, a blow to Berkeley.
- An extension to the European Convention on Human Rights and Biomedicine covering genetic tests come into force on July 1, ten years after it was written. It covers “Article 3 – Primacy of the human being: The interests and welfare of the human being concerned by genetic tests covered by this Protocol shall prevail over the sole interest of society or science, ” and bans genetic discrimination. It outlines the procedures involved if a person is not able to give consent, and clarifies that the information from a test should be made available to the person.
- The NASEM have recommended that researchers should attempt to give back data to research participants, including those not very medically significant.
- A four year effort has culminated in a recommendation not to sequence all newborns, published by the Hastings Center in the context of a whole issue dedicated to investigating the ethics of the issue. Meanwhile, the four year BabySeq project reports a low uptake in the study by parents of newborns — only 10% were interested in an enrollment session, of whom 67% signed up.
- The UK based Nuffield Council on bioethics issued a report saying that human germline editing is not in and off itself ethically problematic, if it is in the best interests of the child and does not exacerbate social inequalities.
- Projects that utilize “Ethics dumping” will no longer receive EU funding — this is the practice where research that would be unethical in one country is performed in another with laxer standards.
- The UN has been meeting about how to regulate the genetic riches of the oceans. Genetic prospectors have filed 13,000 patents, half owned by one chemical manufacturer.
- Genetic editing will be treated as a form of genetic modification in Europe — with large implications for agriculture. Japan may be about to take the opposite route.
- Gene therapy trials have traditionally requires an additional layer of oversight, provided by the Recombinant DNA Advisory Committee (RAC). But that special treatment will no longer needed. Gottlieb and NIH Director Francis Collins wrote “there is no longer sufficient evidence to claim that the risks of gene therapy are entirely unique and unpredictable—or that the field still requires special oversight that falls outside our existing framework for ensuring safety.”
- The UK opposition party has called for a ban on using Non-Invasive prenatal screening for determining sex.
- In the 1970s, scientists converged on a “14 day rule”: embryos could be studied up to 14 days post fertilization, which is when the “first inklings” of a nervous system first appear, and the point at which the embryo can no longer split. At that time, it wasn’t possible to keep embryos alive in cell culture beyond a few days. Now, scientists are just about at the 14 day limit. To push beyond this limit, some researchers have used “synthetic embryo like structures” made from stem-cells. This piece covers the advances.
Science
- Chinese scientists report on base editing of a Marfan syndrome variant in viable human embryos. Base editing is a form of gene editing that does not introduce a double stranded break, and hence should be less prone to error. This is another first for China in this space. 18 of 18 embryos had the desired variant edited, and 2 of 18 had additional off target effects.
- Link between Alzheimers and the herpes virus — based on RNA sequencing from brains
- CRISPR can cause large scale structural variants that may be pathogenic. The news sent companies using the technology stocks’ down.
- Successful gene editing of mice’s brains using a non-viral delivery of CRISPR. The phenotype, repetitive behaviors, reduced in mice models of autism.
- Gene drives in mammals reported for the first time — the inheritance of a specific genetic variant went from 50% to 73% in female embryos only, suggesting the controversial tech has a long way to go. (Note that gene drive research in mosquitos is being actively pursued.)
- A case of uniparental diploidy, where most of an 11 year old’s cells only have DNA from her father. She has not yet presented malignancy, though all of the other ~20 such cases worldwide have presented malignancy early in life.
- An accessible introduction to “Inborn errors of immunity” — where genetics, rather than unspecified bad luck, can lead to someone succumbing to an infection. Knowledge of the underlying genetics can lead to successful therapeutic approaches.
- ExPecto, a tool to predict the effect on gene expression of any genetic variant.
- A tool to predict how many SNPs, and hence what sample sizes, are likely to be involved in various traits.
- Knowledge of how genetics influences behavior can lead to “free will doubt”, with potential implications for how individuals view justice. In general, free will doubt decreases support for justice based on retribution, towards justice based on consequentialism. A study shows that, while this is true for violent crime, for “low affect” crime, free will doubt actually decreases support for either type of justice.
- Evidence that whole genome or exome testing outperforms the current standard of care (chromosomal microarray) for children born with suspected genetic disease.
- The genome can gather mutations as we go through life (a phenomena well known as the cause of cancer). A study (and a nice write-up) shows that early life experiences can end up written in your DNA — some genetic mutations in brain cells depend on the amount of maternal care baby mice received.
- Genes responsible for antibiotic resistance have been located in the air of all major cities tested, with San Francisco topping the list.
- Variable penetrance significantly complicates the picture linking genetics to disease — this is the phenomenon by which the same genetic variant can have a very different effect in different individuals. A new study shows that the variable penetrance of rare coding variants can often be explained by the common regulatory variants for the coding regions in question. (This write-up is more accessible.)
- Your DNA, Your Say, an international online survey about attitudes to genomic data sharing using videos to educate has recruited 20,000+ participants in at least 14 languages.
- The genes that are most studied can be predicted based on a handful of factors indicating how easy they were to study decades ago, rather than on factors which best indicate which are likely to be the most disease relevant. They suggest funding changes are needed to counter this preference for conformity.
And finally, a nice piece from science historian Oren Harman on use of the term “gene” — historically and in the future. And another lovely piece about how some of the genes got their names, including the background to “Pray for Elves”, which I learnt originates with someone I used to work with, Prof Mark Yandell.