The era of fast, cheap genome sequencing is here
wired.com
wired.com
Also how many full sequences would I have to get before the data is statistically sound. Since there is likely a base pair or two that will be different between the cells (maybe) and the sequencer hardware and process will likely have errors too. My gut tells me 10 to 20 but perhaps that seems like overkill since many base-pairs might be junk/repeats/irrelevant anyway and the raw data will have some confidence indication as well IIRC.
There are a number of companies that will do 30x WGS and give you the BAM file. I believe Nebula is one of them, but know there are other names out there.
I’ve yet to finish the order process but I share this here because in my research of companies providing full sequencing services I never came across them and yet they seem far more reputable than others. Current SEO for these services turns up a lot of sketchy companies.
Full body MRI can be had for ~$2k. I can imagine going in and getting a genome sequencing and an MRI scan done, and then over the years having a much better idea of the changes going on in your body as you age, along with an evolving map of what to look out for.
It's exiting, and yet I'm worried about handing over information that could be used against me by corporations, the government or insurance companies.
I know there are efforts like Galleri [1] to detect cancer early using liquid biopsy [2] but I don’t know how effective they are at the moment.
My own father is a physican and very against any elective exams himself. His reasoning is that his colleagues are excessively interventive so the less medicine, the better. I wish your own father had been diagnosed in time but I can see the downside in doing too many exams as well. Doctors won't be able to keep their hands off treating any little lump they find, even if it's to your detriment.
Case in point, surgery for back pain causes said pains to worsen at about the same rate as it improves them.
Of course, with tests that are of themselves dangerous, that's another issue. Skip those if the risk/reward is too high.
Getting diagnosed too late is much worse than having to say no.
I noticed this first when tracking sleep data, but it can also be applied to many other areas were humans try fix small problems but tend to make things worse.
I still think there are probably big obvious things that need to be caught early.
However, the kind of detail you require to catch something small, early requires an MRI--generally with a contrast dye. Unfortunately, Gadolinium, for example, seems to bioaccumulate. So, just like X-Rays, how many MRI's do you need to get over your life before the MRI is more problematic than the thing you are looking for?
And this is before you start getting into what emotional toll that "spotting something" will have on the patient. A friend of mine had "something" on her mammogram and the anxiety combined with high-blood pressure spike that induced until the biopsy came back negative could very well have given her a stroke.
The worst part is that people want to do all this "medical" stuff when there are much better low-hanging fruits: eat better, exercise, sleep, get vaccinated (viruses cause more damage than you think), etc.
We are getting HRR genome testing done for my father this week for his prostate cancer. Some of the PARP inhibitors seem to have decent success in prostate cancer.
They gave my father one year if he went on Chemo and the chemo killed him in there weeks.
Check out these links about the ABO gene and Manganese in pancreatic cancer.
https://pubmed.ncbi.nlm.nih.gov/12649190/
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2657095/
https://pubmed.ncbi.nlm.nih.gov/12700280/
I suffer from a lot of genetic stuff that I have fixed understanding my genetics. That is more valuable than any harm that could be done with my genetics.
But, before you conclude that you want to try whole body MRI as your response, I recommend you read this: https://cancerimagingjournal.biomedcentral.com/articles/10.1.... It's heavy going if your're not au fait with statistics, but important. The potential for more harm than good is quite real.
Though many of the other comments here are focused on the negative consequences of doing full body scans, I believe they are thinking about full body MRI and other tests using only today's technology.
They are correct, we can't or maybe I should say probably shouldn't give a doctor a full MRI and say "what do you see that might need to be addressed", however, we should be able to train an AI to say "look at this scan, is there anything that you would recommend the doctor take a closer look at".
This should probably be done in conjunction with blood work, and maybe even genetic testing, etc.
A doctor looking for something wrong has a TON of things to look at, and interactions between them, what other patients have had issues recently that might lead them in one direction or another due to their experience.
An AI can likely be trained to take a less biased, or perhaps more appropriately biased look at significantly more data than a person, and make recommendations to the expert doctor of what the AI suggests would be beneficial to look at.
It can take incomplete data, like the kind you get from 23andMe, etc, and magically "fill in the gaps" to compute the majority of the rest of your genome
There used to be a free service https://impute.me that would do this for you
https://twitter.com/lassefolkersen/status/154763708634953728...
Search Google Scholar for "Lasse Folkersen", he is active in the field and has a lot of great papers about this
The law will absolutely have to adapt to these changes.
You cannot even reliably determine a donor's hair color, from a single sperm cell.
I imagine it's much more difficult to reconstruct a full genome from a sperm donation than from, say, saliva on a discarded cup.
A sperm cell contains one complete haploid genome.
But (my bio is rusty) a haploid genome (gamete) cannot fully represent the genome of a diploid human.
If you're looking for genetic disorders, these will often be the result of recessive traits, right? So having only half of the chromosome sets is not indicative.
That's my example of hair color. With a full genome sequence, you could determine the person's hair color, but with a haploid genome, you could not.
If OP's joke is about the privacy implications of sperm donation, then saliva seems more risky.
* This I think indicates some other genetic and/or environmental factors we don't have a good grasp of.
Correct; my broader point was that when most people refer to a "human genome" it is ambiguous whether they mean diploid or haploid unless specified. Metrics like "3 billion base pairs" etc. actually describe the haploid human genome.
> If you're looking for genetic disorders, these will often be the result of recessive traits, right? So having only half of the chromosome sets is not indicative.
Often, but not always; there are myriad autosomal dominant disorders, as well as sex-linked disorders which require only a single mutant allele on X
> If OP's joke is about the privacy implications of sperm donation, then saliva seems more risky.
Here I think both with STR and SNP profiling you are equally at risk (i.e., saliva is not more risky)
Think about it. Each sperm cell gets a random mix of your chromosome sets. There are 2^23 possibilities.
But of course you only need to find the matched pair for each chromosome. So you'd only need a few hundred cells.
I'm not a geneticist, but that sounds like a much harder genome sequencing/assembly problem than the fully matched set you'd get from saliva etc.
Are you worried that companies will not respect this law? Or are you not US based.
You can't sign a contract to sign yourself into slavery, it's just straight out illegal. And thus you don't see anybody being signed into slavery via contract in the USA, and there many other things that act this way in law.
Because it is highly regulated, with no known examples of insurance companies and regulators colluding to incorporate health data for pricing purposes.
https://www.healthcare.gov/how-plans-set-your-premiums
For example, you can even figure out your health insurance premium in the state of NJ just based on this pdf:
https://www.state.nj.us/dobi/division_insurance/ihcseh/ihcra...
https://www.healthcare.gov/how-plans-set-your-premiums
> FYI Your health, medical history, or gender can’t affect your premium.
However, life insurance, long term disability insurance, and I presume other types of insurance are free to use health data for pricing. Of course, if genetic data was known to be worth it for predicting probability of death or disability by a certain age, those insurance companies would already be requiring applicants to provide genetic data.
I'm not telling other people they are wrong to be afraid, but I want to understand the exact mechanism of those fears.
And if you're in another country, especially european ones, look into other research programs, there are lots round
Beyond just eating more health (in whatever way) and exercising more, which does go a long way. Thse drugs are trying to be extremely targeted and intricate in now they work
For example https://fjfsdata01prod.blob.core.windows.net/articles/files/...
Human nature values first impressions and you can expose your flaws after you have established your reputation a bit. This data has the potential to short-circuit that and could lead to yet another technology that is maladapted to human nature.
if/when the time comes of a meaningful societal regression like this, i want it to not be so easy to drag me into it. it’s precautionary.
Also the cancer drugs like PARP inhibitors seem to be helping everyone regardless of mutation status in clinical trials. So the connection is much less clear.
What I was saying was that those drugs have been proving effective regardless of mutation status. So mutation x = give drug Y connection is not really well understood.
Example, olaprib for Prostate cancer
> the addition of olaparib was associated with a significantly prolonged radiographic PFS irrespective of HRR status
https://www.practiceupdate.com/content/asco-gu-2022-propel-s...
I don't believe this is the case in general. Indeed, all cancer drugs fail sometimes, and there has to be something different about the cancers they fail on. If not the mutation status, then what?
Even FDA is approving these mutation targeting drugs without mutation status. Niraparib was approved for Ovarian Cancer even if you don't have HRD mutations. Approval was based on NOVA trial, you can lookup the details.
I am not saying there isn't any connection, what I am saying is it not really well understood or clearcut as mutation x = drug y that general public believes about these drugs.
We’re starting to get some good utility from sequencing. Currently whole exome is more informative than whole genome. But, there are definitely some good clinical uses!
https://rarediseases.info.nih.gov/diseases/9146/cardiofacioc...
Just because doctors do not know what to do with it (they are decades behind most research) it does not mean you cannot learn about this and take action on your own.
Just knowing my FADS1, FADS2 genetics was enough to change my diet to seafood only and raise my HDL from a consistent 35 to over 55 in three weeks. Doctors put me on a statin which lowered my HDL to 25.
Seafood helps everyone with HDL. Is 35 -> 55 not possible with those genes that you mentioned ?
Having a particular gene doesn't mean that it's actually turned on. Your heart and your hair cells have the exact same genome, but do vastly different things.
I am not excited as I used to be about the promise of cheap genome sequencing by itself. It needs to be paired with lots of other data for it to have any use, except for determining phylogeny/ancestry.
Even more exciting than this would be cheap DNA synthesis, allowing us to build and test complicated genetic circuits in fewer steps. The current methods usually rely on stitching together existing sequences, rather than creating what you need from scratch.
The price of DNA sequencing has fallen much faster than the price of DNA synthesis. Here is a comparison of the cost of DNA sequencing vs synthesis: http://www.synthesis.cc/synthesis/2016/03/on_dna_and_transis.... It would be good to find more recent data though.
Here is the archive of the wired article, btw: https://web.archive.org/web/20221001173347/https://www.wired...
*To produce the raw data, and does not include analysis
It's likely the bill for a consultant (senior doctor) to examine a patient, order a test then read and sign off on a report is already more than the chemistry + cpu + analysis already.