Big themes in medical genomics - ASHG2021 overview
For a second year in a row, Merogenomics had the pleasure to attend the latest conference on human genetics put on by the American Society of Human Genetics. It is the largest gathering of scientists and industry players in one space with a presentation of the current state of research in human genomics, which is especially valuable for demonstrating developments with exciting potential medical impact. In this post we summarize what Merogenomics perceived as the biggest themes. We also mention what themes are emerging, what is the usual information and what did not get as much attention as we might have expected.
Let’s get started with the biggest themes!
Conference’s biggest themes
Polygenic risk scores
Typically we rely on a family history to estimate the risk of the development of specific condition in a lifetime. However, another fantastic tool that is emerging for many conditions is what is referred to as polygenic risk scores. Polygenic risk, or a cumulative disease risk stemming from the influence of many different genetic factors, can be the equivalent to a monogenic mutation impact. Monogenic mutations are specific mutations affecting one gene but they are so disruptive to proper biological function that they confer a high risk of disease development. Monogenic diseases have been, thus far, the main focus of genomics studies due to the difficulty of estimating the polygenic risk, or the risk resulting from many mutations in unison. However, now that genomic studies have been performed on an enormous number of people with detailed health records, the development of polygenic risk scores is constantly increasing and likely will be entering common clinical practice in the not too distance future.
Coronary heart disease polygenic risk is probably one of the earliest such examples to be demonstrated and has since been improved, and we certainly had a glimpse of that in spades at the ASHG. But many other such risk scores have been demonstrated that we have not seen before. One interesting example was a polygenic risk score for epilepsies. To show you how these risk scores work, in that study the authors showed that those scoring in the top 10% of the polygenic risk score for epilepsies based on their personal genetics had more than double the lifetime risk of experiencing epilepsies than those in the remaining 90%. It is such fine tuning of the predictive outcomes for individual patients that is so exciting for the future of personalized medicine. And this is on the low end of some of the best performing polygenic risk scores so far developed. Others include prostate and breast cancer (contact Merogenomics for access to clinical use of such polygenic risk scores for unaffected family members with history of cancer), atrial fibrillation, type 2 diabetes, obesity, schizophrenia, Alzheimer’s disease, bipolar disorder, hereditary hemochromatosis, and arthritis, just to name a few. One group from Finland in fact showed that polygenic risk scores can be used independently of family history and that family history can rarely outperform polygenic risk scores. Definitely exciting developments for doctors.
Merging omics data and databases
The continuing theme of the merging of different “omics” fields to increase the comprehension of how molecular biology can drive the development of different traits, including health outcomes, appears to be gaining momentum. To illustrate that, we are already very familiar with the sequencing of germline DNA (the DNA you are born with) for possible mutations involved in health (by the way, mutations are instead now referred to as variants in the sciences). These can now be complimented with a greater comprehension of structural variants (mutations involving a very large segment of DNA) which is not normally deciphered easily with short read DNA sequencing technology (the predominant technology used in decoding DNA sequences). You can add to that somatic variants (mutations in DNA happening post-birth), transcriptome (analysis of which RNAs are produced in a given cell line to determine which genes are being used, as well what potential mutations could arise in the production of RNA) and finally epigenetics (decoration of DNA with specific chemical which can influence how the DNA is used to produce specific RNAs). And this is just the field of genomics! On top of that one can analyze the proteomics (determing which proteins are produced from RNAs in a given cell line), metabolomics (determining which chemical compounds are produced in a given tissue) and even what other organisms inhabiting our body (microbiome) which might be influenced by all these molecular events, and in turn also influence them. The point is that there can be a myriad of influencing factors behind a specific biological outcome and we are diving deeper and deeper in understanding how it all might work in unison. But we are still just cracking the door to this vast universe.
Granularity increasing – more comprehension of utility
This is undoubtedly one of the more exciting areas of research where the quantity of data is finally allowing a much higher level of prediction of possible outcomes. One of the most obvious questions anyone undergoing predictive genetic testing and obtaining a pathogenic result, is to ask if this will result in a disease. This is one of the most complicated tasks of medical genetics - being able to predict the probability of developing a particular disease given a particular mutation (variant). Such probability is referred to as disease penetrance, and a few studies dedicated to unravelling this topic were presented, including estimations of penetrance not only for specific genes, but now even down to specific individual variants. This will be one of the more valuable tools available to doctors in the future as medical DNA testing continues to expand, as it is one area in which doctors struggle. Another valuable example is being able to link more symptoms to specific genes. One large study using UK Biobank data showed that 69 genes were linked to nearly 6 traits per gene! When a single gene or variant is responsible for multiple different traits, this is referred to as pleiotropy. This type of information will be very valuable in the future in being able to link traits to genes and then more rapidly identify the genetic roots of undiagnosed conditions.
COVID-19 and immunology
Studies related to COVID-19 disease were prominent at the conference, not surprisingly. One interesting investigation presented by a group from the University of Surrey was a confirmation of genetic association for a previously suspected link between type 2 diabetes and reduced lung function, which could help explain why such patients have increased risk of death post SARS-CoV-2 infection.
Another study that involved full SARS-CoV-2 genome sequencing of 25K COVID-19 positive samples indicated an interesting correlation between quality genome detection and the PCR cycle threshold (Ct) values, once again showing that high PCR Ct values might be problematic for accurate diagnosis. Typically, no quality genome was identified past Ct of 28. Furthermore, small portions of the PCR positive samples could not be confirmed at all with genome sequencing, and the average Ct value of these presumed false negatives was still only Ct of 24 while the average Ct for all remaining true positives was only at Ct of 22.
Another interesting presentation by a group from University of Tampa studied the distribution of variants involved in increased expression of ACE2, the receptor used by SARS-CoV-2 to invade our cells. It appears that more than 50% of males around the world have variants that enhance ACE2 production, and they are predominantly seen in the countries most hit by COVID-19. Males also appear to have a higher expression of ACE2 than females. This is because the ACE2 gene is on the X-chromosome and in women, one of the X chromosomes is silenced, meaning it will not be used. Thus there is 50% chance that the X chromosome with variants predisposing for higher production of ACE2 will be silenced in women. The take home message of the authors was that genetics might be important in analysis for how to successfully fight COVID-19. Can you imagine if your genome sequencing was mandated for this purpose?
One of our favourite studies was one that analyzed the epigenetic differences between healthy and COVID-19 patients for any strong statistical association between the epigenetic status and health outcome. Just like the GWAS studies we have often written about in the past, but this is referred to as Epigenome-Wide Association Studies (EWAS). In this way, it was shown that infected patients exhibited epigenetic modification in genes related in viral infection response, which makes sense, but more imporantly this information was then demonstrated that it could be used to determine who is infected and then predict potential worsening of clinical outcomes.
Now let’s jump into the emerging developments of what we might have the potential for great impacts in future medicine.
Emerging influential development
Single cell and spatial genomics
This is definitely one of the more exciting evolutions in genomics technologies, where the genetics or protein make up of individual cells in a tissue can be probed independently at a same time. One way to think of this technology is like trying to compare an X-ray with CT scan. Just the scale of additional detail obtained is enormous. Currently up to 20K cells can be analyzed individually from a given tissue. Alternatively, same samples could be analyzed for specific reaction to drug candiates or any other agent that we might be curious of how it imparts its change on the sample. These cells or samples can separately be screened for genetic information, RNA production, protein make up, thus allowing very precise picture to emerge as to what is going on biologically. This information can be surveyed at specific regions of interest inside the tissue which can be defined by the user (for example, based on histopathological staining of the sample, or protein expression patterns), either to probe specific biological compartments , specific unique cells or to probe the contours of one specific biological area as it gradually changes into another (thus helping to understand what governs the diffusion pattern). Alternatively, entire tissue can be analyzed in a grid manner with subsequent analysis for details. Really exciting area of innovation, especially for cancer pathology, and we discussed single cell and spatial genomics recently based on review of another scientific summit.
Antisense oligonucleotide therapy
One very exciting area of development that was presented on is a new method of treatment for genetic diseases that is currently being investigated. It is antisense oligonucleotide therapy, or building synthetic, short fragments of the RNA of a specific sequence which can be used to bind to mRNA (produced inside our nucleus to act as a template to build proteins inside the cell). The goal behind this approach is to bind and prevent specific mRNAs from being used by the patient’s body by actually binding the mRNA with complimentary RNA and thus making the mRNA unavailable for use (hence calling this approach “antisense” where the mRNA would be the “sense” template). This could be used in instances where the disease is produced as a consequence of production of too much of a given protein leading to a disease. Alternatively, it could be used to diminish the production of mutant protein if that protein is produced from a mutated gene template, but if a second, normal unaffected gene template is also available for production of a normal unaffected protein, then reducing the production of the mutant protein could ameleorate the disease. In the media this novel approach has been aptly nicknamed “programmable medicine”. The first example of this antisense oligonucleotide therapy has only recently been approved for spinal muscular atrophy, but currently other conditions are being investigated with this treatment approach.
While still too early to celebrate, this has the potential to be an incredibly useful tool for treating rare diseases that otherwise might not see the development of treatments (because they are too rare). When grouped together though, these thousands of different rare diseases in the population afflict a substantial segment of our society and thus this approach could be very exciting and important to the future of individualized medicine.
CfDNA for clinical use
Circulating cell-free DNA (cfDNA) is not a new technology in medicine. It is already employed in such applications as: non-invasive prenatal testing when screening fetal development for potential chromosomal abormalities; also as a liquid biopsy to screen for cancer development; and finally for tracking organ transplant rejection. But, it is the new, additional acknowledgement of what else we can learn from these circulating short fragments of DNA that is the exciting area of development.
For example, a team from Finland presented new information on how tracking cfDNA in serum in cancer patients - with an initially good prognosis but who later developed a metastatic condition - was sufficient to identify important mutations in the primary tumour and any potential subsequent recurrences prior to clinical diagnosis. This obviously looks like a very promising tool for cancer patients and cancer survivors to have in their toolkit to fight the disease. The authors of this work previously noted that the mutations found in cfDNA might also be different than those found in biopsies, indicating the need to sequence both these sources of genetic information, but overall just the increased amount of circulating tumour DNA appears to be a strong indicator of relapse and the subsequent risk of a poor recovery.
One of the more exciting developments presented by the research group of the Ontario Institute for Cancer Research (and others) showed that cfDNA from cancer can be tracked for epigenetic changes allowing for tissue origin and this can be observed for many years prior to an actual cancer diagnosis based on clinical symptoms. Most importantly this information can be used to predict early cancer development so mitigating action can be taken rapidly. Wow! Such fantastic developments and we are so looking forward to this technology taking a foothold in regular health care in the future.
There are other frontiers being broken with cfDNA, where the size of the released DNA fragments can give us information about the person including: age, health status or even gene expression, but this is still in the early stages of translation for medical use. A greater depth of understanding and research is required to fully grasp the many opportunities of what can we learn from these fragmented pieces of DNA.
We wondered if we could pick one presentation to feature in this article. There is simply too many choices. But we got one because it encompasses many of the interesting topics already discussed. Merogenomics top pick was a presentation on ethical and social implications of clinical use polygenic risk scores (we told you it is coming to a doctor near you!). It was a panel of women discussing research into opinions and attitudes about the safe use of polygenic risk scores as expressed by both patients and doctors. These groups have different goals because patient is usually driven by curiosity for information that they feel is relevant to them, whereas doctors are primarily driven by the desire to minimize harm to patient. That means very contrasting opinions can be produced depending who your talk to (and perhaps why patients and doctors can come into conflicting conclusions).
The other very interesting component briefly discussed is that segregation of our population by race is problematic if we are attempting to capture potential genetic differences influencing health outcomes. That is because apparently it does not reflect accurate information. The more accurate way to link ethnic differences to genetics is to use continental divisions, as that is more accurate reflection of differentiating population for specific genetic differences to better administer medicine.
Thus the more appropriate way of dividing people into clusters would be based on Africa, Europe, Asia, Oceania, and the Americas. In other words, for best medical outcomes, we should be thinking of ourselves not in terms of race, but in terms of continents of origin.
This is a novel concept in genetics, since use of race is so pervasive in sciences to group people, but evolutionarily it makes sense. Any barriers that significantly inhibited human travel would have led to accumulation of population specific variants. We divided world map into continents based on grand geographical divides and such divides would limit easy mixing of populations. Within continents there could be further isolation of populations leading to unique genetic patterns. Geographical isolation of populations was in fact common due to limitation in transport. Some populations are very famous for geographical uniqueness in the world of genetics, and are studied in great detail to learn about specific medical conditions. But the benefits of such studies are distributed to the entire world. The Finnish population is one such example.
Where greater blending of people was allowed to occur, the more distributed the genetics became as well. These instances create different maps of human genome, and the entire world could be mapped in this manner into one giant family tree. But the differences can also be fantastic in nature. We are yet to learn true shape of human genomes on a global scale. With the completion of the human genomes thanks to long read technology, discussed at the conference, we are going to learn how diverse the chromosomes are going to be between different corners of the world. It is going to be fun adventure learning about these unique divisions among one enormous family tree. It is a weird concept to grasp for us humans who like to seek distinction, but then remember, all of the SARS-CoV-2 viruses infecting the entire world is a giant family tree of related members. Humans are also one giant family, and we have our own unique strains. Unique flavours sound nicer though. Just to differentiate us a bit from the virus.
If you were to paint each continent as a specific color, and map human genomes only based on those colors, you would still end up with a rainbow of multiple continents, due to mixing of humanity, and the fact that Africa gave birth to this family tree, with the global family evolving from those ancestral roots. Certain continents can be represented more in some individuals than others. Each continent of our genomes has its own medical record to deal with. Currently we predominantly studied European continent the most in medical genomics. But as mentioned, all ancestral groups are currently investigated precisely because it is understood that this is one of the fastest way to learn genetic diversity affecting medical outcomes, as such mutations of medical impact could be observed randomly anywhere in the world. There is more to it. Diversity allows greater discovery related to identifying treatment. So it is wonderful to see how rapidly the gaps are being filled. Enhanced diversity of genomics was another big theme of the conference. Especially Africa has gained significant grounds in representation in genomics, and instinctively that is smart, because that is our ancestral group. It is tied to the entire world. What we discover in Africa continental group of people should have an impact on greater understanding of the rest of the entire population of world. However, this representation is still behind other ethnicities and has great room to grow further to realize these gains.
What you always see presented in enormous quantity at conferences such as these are discoveries of new variants linked to different traits. This of course does not mean they are valid - that will take some time to ascertain - but proposed associations between specific variants and human traits are constantly being uncovered. Validation of certain variants however, is now also increasing thanks to the continued amount of medical cases solved by medical genetics, which is also presented in a copious amounts at conferences. It always feels like every disease under the sun gets presented (ok, a mild exageration, considering that many thousands of actually conditions exist).
Another common trend is the expanding use of large data sets. Studies involving hundreds of thousands of individuals, or even millions of individuals are not uncommon now. The Global Biobank Meta-analysis Initiative introduced at the conference is a perfect example. It is a consortim of international databases dedicated to collecting the genomes and traits of the participants. Together, this international effort has assembled 2.6 million samples across diverse ancestries. The goal is to gather information correlating DNA variation to traits, including human diseases. Unfortunately, Africa is underrepresented. Perhaps the world should consider coming together to invest in Africa’s scientists to build their own large scale, pan-continent database because study after study indicates how richly diverse the genomic architecture is in the population of Africa.
Next some segments of research that Merogenomics would perhaps expect to be presented in greater quantity than what we managed to find.
Left in the dust
NIPT and prenatal testing
One significant deficit of information compared to past conferences was in research related to non-invasive “screening” (NIPS), more frequently referred to as “testing” (NIPT). This is a big areas of interest for Merogenomics so we were looking out for this. One notable exception though was research from the Netherlands (not a surprise, as the Netherlands has been an undisputed world leader in NIPT research due to being one of the earliest adopters of this technology as part of their standard care). In some fascinating work from Amsterdam UMC was a demonstraton of the use of NIPT to track viral infections in pregnant women! This is a fantastic development and it is potentially, extremely valuable information to be able to tracked during pregnancy because viral infections during pregnancy can be of serious consequences to both the mother and the baby. We can’t wait to cover this in more detail!
More new tech
Looking out for new technology is always exciting. We did not see much covered by independent scientists. The one notable and very promising exception was information presented on the expanded capabilities of optical genome maping with Bionano technology. Normally this technology stretches very long pieces of DNA (more than 150K bases! Just cosmic proportions) and labels it with a dye in a specific spots based on where the DNA is nicked by a restriction enzyme. In the process of fixing the nick, the dye is incorporated and all the restriction sites on long stretches of DNA are labelled. Why this works so effectively is because we can easily identify where these restriction sites are in the genome. The result is the ability to study structural variants of any kind that otherwise might not be uncovered by short read sequencing. However, in addition to this neat trick, now Bionano can also add epigenetic analysis where once again specific sites are labelled with a different colour dye. This allows the study of both genetic and epigenetic events in a single assay! On top of that, this can be done in a haplotype specific manner, allowing the determination of which regions of the genome were inherited from either mother or father. We were pretty excited about this development.
Clinical utility studies
We also love to see clinical utility studies that help empower the use of genomics in medicine. This time around there was not much to be seen. There was a presentation on the results of SouthSeq study, a study involving numerous medical centers in the south-western US, which investigated genome sequencing as a first-line of genetic testing in newborns with complications. After all, as per other content presented at the conference, 2.5% of newborns require intensive care due to undiagnosed diseases. The conclusion was typical of what we have seen numerous times in the past, that genome sequencing is more effective in diagnosing conditions than the typical standard approach of using a battery of individual clinical genetic or biochemical tests. An additional advantage of genome sequencing is that no prior assumptions have to be made as to which genes need to be inspected as all genetic information is inspected in one go.
Another study from Harvard that showed screening newborns at birth for cancer predisposition would result in cost-savings for the majority of cancer predisposing genes due to a subsequent cascade of testing of family members that otherwise are unaware of their predisposition and disease risk. We don’t doubt that newborn screening with genome sequencing will be a norm everywhere one day.
Overall, as always, the American Society of Human Genetics 2021 conference was such an insanely informative one and of such massive proportions that it was difficult to get through all of the content. Luckily the online version actually made it easier to be able to sift through so much information, and to generate a pretty good overall picture of the current state of research in human genetics. This article tried to summarize the biggest, most exciting trends but so much of the research deserves closer scrutiny due to many provocative developments. We definitely hope to bring you some of it in the near future!
This article has been produced by Merogenomics Inc. and edited by Jason Chouinard, B.Sc. Reproduction and reuse of any portion of this content requires Merogenomics Inc. permission and source acknowledgment. It is your responsibility to obtain additional permissions from the third party owners that might be cited by Merogenomics Inc. Merogenomics Inc. disclaims any responsibility for any use you make of content owned by third parties without their permission.
Products and Services Promoted by Merogenomics Inc.
Select target group for DNA testing
Or select popular DNA test
Pharmaco-genetic gene panel
Non-invasive prenatal screening
Cancer predisposition gene panel