The Sequence 5/8-5/14
The Human Pangenome is Here, Genetic Testing Included in Employee Benefits, New Understanding of the Interaction of Genes Expressed in Schizophrenia, AI for Genetic Counseling
Nature. Volume 617 Issue 7960, 11 May 2023; Image credit: National Human Genome Research Institute
The human pangenome is here
Back in February, I wrote a newsletter article discussing the draft of the human pangenome; an updated version of the human reference genome that is made up of 47 genomes. See that article and learn more about what the human reference genome is here. Fast forward to this week, when the human pangenome was officially published.
Can you give me a recap of the original story?
The current human reference genome is based on the DNA from just a few people from Buffalo, NY. A problem, because it introduced biases in how patient sequences are interpreted.
Why does it introduce biases?
During exome or genome sequencing, patient sequences are first aligned to the reference genome sequence, and then reduced to a set of differences from that reference. So, clearly, it matters what the reference looks like. And historically, the reference has not been diverse nor representative of very much structural complexity.
The construction of the human pangenome; Image credit: Describing genetic diversity with graphs by Arya Massarat and Melissa Gymrek
That’s a problem.
It is. And that is why the pangenome is so important. Not only does it contain a more diverse population set, it is made up of a collection of 47 genomes as opposed to single individual genomes. The authors used three graph-building methods to construct pangenomes from these 47 assemblies, creating a pangenome representing many genomes from different ancestries. I think Eric Green, director of the US National Human Genome Research Institute, said it best in pointing out that the pangenome is crucial in advancing the field of genomics in an equitable way.
Having a more accurate picture of genetic variation across the population can tell us a lot about both our current health, as well as inform us about our ancestors as we leverage our ability to analyze DNA from historic objects. Interested in learning more about articles uncovered from the past? This publication describes the genetic diversity and range of ancestry groups present on the Mongolian steppe thousands of years ago, this one talks about what scientists learned from analyzing DNA from a deer tooth pendant, and this one talks about how a genetic variant originating in Neanderthal DNA influences nose shape.
Genetic testing included in employee benefits
Everest, a global provider of reinsurance and insurance products, first decided to incorporate genetic testing into its comprehensive benefits portfolio in 2018. By offering genetic testing at a reduced cost for employees, the company not only encouraged employees to be proactive about their health, but prospectively cut down on healthcare costs. Here, I summarize an article by Human Resource Executive on the results the company has seen since deciding to cover the genetic test for employees.
How does that cost-savings work?
Cancer treatments were the top healthcare cost for employers in 2022. To that, add that ~45% of cancers are linked to genetic or modifiable factors, meaning the cancers can be prevented, or at the very least can usually be detected earlier. So, if employees can be proactive about cancer screening and treatment, the costs associated with treatment will go down. According to the article, early-treatment costs can be 50%-90% less than treatment at later stages.
An astounding 10% of Everest employees who took advantage of the genetic testing were found to be at increased risk for cancer due to either genetic variation or family history, and eight cancers were discovered as a result of the testing. Considering survival rates of cancer are higher when it is detected earlier, it sounds like it was worth it.
What’s the takeaway?
To me, this is an insurance company acting in integrity and practicing what they preach. In covering health costs for employees, it certainly makes sense to cover testing for cancer; especially when nearly 40% of Americans will develop some type of cancer in their lifetimes.
Finally, in light of the publication of the human pangenome, I’m inclined to point out the diversity and inclusion component to this benefit. By making this kind of testing easily accessible through work health insurance, employees take an important part in increasing access to health testing and preventive health.
New understanding of the interaction of genes expressed in schizophrenia
Pergola et al. studied the expression levels of different schizophrenia-associated genes that were previously identified using genome-wide association studies (GWAS; see my post explaining GWAS here) in 833 post-mortem brain samples. These included 562 brains from neurotypical controls from European and African American individuals, and 186 brains from schizophrenia patients. The purpose was to identify whether there are certain changes in the DNA that are expressed differently in people with different manifestations of schizophrenia.
The results?
Interestingly, genes associated with schizophrenia were expressed differently depending on the age of the individuals. In separating samples by age, the researchers were able to get a better idea of which genes were really being expressed and impacting the brain function of the patients.
Interesting. Is any of this different from what we already knew?
Yes! We already knew many, hundreds even, of genes associated with schizophrenia in some capacity. However, it has been largely unclear how these genes affect brain functioning to manifest in symptoms of schizophrenia.
What’s the takeaway here?
Pergola and colleagues found that the expression of various genes involved in schizophrenia changes with age and also differs between brain regions. Important, because this helps explain why different symptoms are onset at different times in life for different individuals. Although genes associated with schizophrenia have been identified in the last decade or so, this research takes us one step closer to understanding why and when those genes have an impact.
In this related article, researcher describe findings from a genome-wide association study focused on finding variants involved in suicidal ideation in Million Veteran Program (MVP) study participants.
AI for genetic counseling
Last week I discussed whether we’re ready for elective genetic testing, or genetic testing that people can order at home. Although the benefits are numerous, the takeaway was that there are a limited number of healthcare providers available to go over results with people who order genetic testing on their own. I also talked about this in this post, where I discussed the challenges associated with population genetic testing, or genetic testing on the entire population for some of the most common genetic disorders; not just people at a higher risk of a genetic disorder.
One of the ways to overcome the challenge of under-staffing would be to use artificial intelligence (AI) to supplement healthcare provider sources. Here, I talk about the results of a study by Koerner et al, which measured the success of a novel delivery mode of genetic counseling: chat.
Interesting. How?
Koerner and colleagues utilized data from DNAvisit, a telehealth company that provides messaging-based genetic counseling.
Where does the AI come in? After analyzing patient inquiries via chat, the AI tool suggested replies for the genetic counselors to send based on a library of prewritten responses to common questions. All of the responses could then be modified by the genetic counselors.
To assess the efficiency, Koerner’s team analyzed chat transcripts from sessions with 34 patients who had received direct-to-consumer genetic test (DTC) results.
After looking at factors such what the patient inquiries were and their satisfaction with the answers they received, the authors write:
"…genetic counselors are able to provide genetic information to clients and respond to their mental and emotional needs through asynchronous chat following DTC-[genetic testing]”
What’s the takeaway?
Although this type of AI is still uncommon in genetic counseling, it could certainly increase access to care for individuals who do not have easy access to a genetic counselor. In addition increasing access, a system like this increases flexibility: patients can ask questions when they want to, even long after they’ve received a result. This is by no means a way to replace genetic counselors, but a way to increase access to information for patients, and a way to lighten the workload of a professional community that is already struggling to keep up with demand.
Interested in more articles about AI in medicine? This article discusses a study testing the abilities of blinded human reviewers to discern human-written scientific abstracts vs. ChatGPT-written abstracts.
And this article walks you through the process of AI training… using Harry Potter books!