Do you have adblock enabled?
 
If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(Boston.com)   Whole genome sequencing offers new way to get money from health insurance companies   (boston.com ) divider line
    More: PSA, genome sequencing, partners, personalized medicine, Greater Boston, Illumina Inc., teaching hospitals, Novartis, genetic tests  
•       •       •

1687 clicks; posted to Business » on 05 Dec 2012 at 11:11 AM (3 years ago)   |   Favorite    |   share:  Share on Twitter share via Email Share on Facebook   more»



27 Comments   (+0 »)
View Voting Results: Smartest and Funniest

Archived thread
 
2012-12-05 11:15:35 AM  
Keep living in the past, Subby. This is awesome for many reasons. Identifying particular genetic variants in your system leads to personalized medicine. Want a drug that cures your cancer without harming any of your normal cells? Have a mutation that results in inactivation of a protein that, on its own is fine, but means that you'll have horrible side effects if you take some drug that the rest of the population is fine with?
I have a client in this space, and they're doing some amazing things.
 
2012-12-05 11:23:10 AM  
You won't get a cancer drug.
Your genetic code shows things we have interpreted as being cancer likely. You're too high a risk.
Your genetic code shows things we have interpreted as causing a short life span. You're not worth it.

/Sure there's good stuff that could be done.
//Way more likely bad stuff.
 
2012-12-05 11:26:27 AM  

ds615: You won't get a cancer drug.
Your genetic code shows things we have interpreted as being cancer likely. You're too high a risk.
Your genetic code shows things we have interpreted as causing a short life span. You're not worth it.

/Sure there's good stuff that could be done.
//Way more likely bad stuff.


You do get a cancer drug.
Aw, shiat. You've got a mutation that makes Erlotinib lethal.
Uh... Sorry?

/guess you should have gotten your genome sequenced
 
2012-12-05 12:28:21 PM  

ds615: You won't get a cancer drug.
Your genetic code shows things we have interpreted as being cancer likely. You're too high a risk.
Your genetic code shows things we have interpreted as causing a short life span. You're not worth it.


Nope, that's illegal.

Under the Genetic Information Nondiscrimination Act of 2008health insurance companies cannot use your genetic information to deny coverage (it also prevents firing, promotion decisions or hiring decisions based on genetic information as well).
 
2012-12-05 12:43:20 PM  
Dr. Lamar: Jerome, never shy, pisses on command. Beautiful piece of equipment you've got there Jerome. I ever told you that?
Vincent: Only every time I'm in here.
Dr. Lamar: Occupational hazard. I see a great many on the course of any given day. Your's just happens to be an exceptional example. Don't know why my folks didn't order one like that for me.
 
2012-12-05 01:44:13 PM  
Hey, I used to work for that lab doing their IT work up until several months ago! :)

I feel bad for the poor schmuck who has to deal with that storage now.
 
2012-12-05 02:48:04 PM  

ds615: Your genetic code shows things we have interpreted as being cancer likely. You're too high a risk.
Your genetic code shows things we have interpreted as causing a short life span. You're not worth it.


I'll take "things that are illegal under the Genetic Information Nondiscrimination Act" for $400.
 
2012-12-05 07:11:05 PM  
As someone who does analysis of whole-genome data, both sequence and microarray,and writes statistical programs for analysis, I'm calling that expensive bullshiat.

There are no validated genetic tests I know of that need microarrays or whole genome sequencing. Nothing done with those platforms on a select onsumer group will yield any information as useful as validated assays already available.

The scope of the problem inherent in big data analysis on a single patient is enormous. We all have many millions of variations, including a proportion of copy number insertion and deletions. What particular variation causes what condition in a single patient? And don't even think that sequencing a single patients cancer tissue is going to get more useful info. There's too much info. How to cull the chaff? Nobody knows.

And don't tell me that proprietary software will help...I write the stuff and the garbage-in rule holds. We know so little about biological systems at the gene and molecular level that a causative aviation or groups of such and their effects are more likely than not to be overlooked because we are unaware of most of them.

But, if someone will pay 9k to fund the bullshiat, why not.
 
2012-12-05 07:14:52 PM  

DO NOT WANT Poster Girl:

And don't tell me that proprietary software will help...I write the stuff and the garbage-in rule holds. We know so little about biological systems at the gene and molecular level that a causative aviation

variation

Ftfm
 
2012-12-05 07:20:51 PM  
DONOTWANTposter girl....I call bullshiat on your bullshiat. Microarray testing is a first tier test for MCA, MR and autism. Although whole genome sequencing isn't completely ready for prime time, the ideal test will be to perform deep sequencing of a panel of genes for actionable mutations and lower level WGS to capture structural rearrangements in cancers.
 
2012-12-05 07:32:06 PM  

DO NOT WANT Poster Girl: The scope of the problem inherent in big data analysis on a single patient is enormous. We all have many millions of variations, including a proportion of copy number insertion and deletions. What particular variation causes what condition in a single patient? And don't even think that sequencing a single patients cancer tissue is going to get more useful info. There's too much info. How to cull the chaff? Nobody knows I don't know.


Good thing you're not everybody then, eh?
 
2012-12-05 07:48:41 PM  

Theaetetus: DO NOT WANT Poster Girl: The scope of the problem inherent in big data analysis on a single patient is enormous. We all have many millions of variations, including a proportion of copy number insertion and deletions. What particular variation causes what condition in a single patient? And don't even think that sequencing a single patients cancer tissue is going to get more useful info. There's too much info. How to cull the chaff? Nobody knows I don't know.

Good thing you're not everybody then, eh?


Sorry, but the biggest challenge in WGS is to make sense of several million variations in light of complex phenotypes. Are you someone actually working in this field ?
 
2012-12-05 08:08:24 PM  

jengen: DONOTWANTposter girl....I call bullshiat on your bullshiat. Microarray testing is a first tier test for MCA, MR and autism. Although whole genome sequencing isn't completely ready for prime time, the ideal test will be to perform deep sequencing of a panel of genes for actionable mutations and lower level WGS to capture structural rearrangements in cancers.


You're full of it. There are no validated DNA microarray assays for a genetic disease. I do work in pre implantation genetic diagnosis analysis, and while we can develop a microarray assay for searching for gross chromosomal abnormalities, it's on a per case basis. Microarrays are not used to diagnose or treat genetic causes of complex disease across patients in a population...anyone who will pay... like the article is suggesting they'll be attempting.

And Autism? They don't even have a cause, just suggestion of copy number variation . And in that,

those slackers at ACMG write, about microarrays for use for that analysis in 2011,

The introduction of genomic microarrays into clinical laboratories has presented unique validation and regulatory challenges. It is not feasible to validate a molecular technology that targets the entire genome in the same manner as an assay that targets a specific gene or syndromic region. Because this technology represents such a tremendous advance in diagnostic utility and patient care, we must accept these limitations and seek a new paradigm of validation and regulation.
 
2012-12-05 08:30:48 PM  

jengen: DONOTWANTposter girl....I call bullshiat on your bullshiat. Microarray testing is a first tier test for MCA, MR and autism. Although whole genome sequencing isn't completely ready for prime time, the ideal test will be to perform deep sequencing of a panel of genes for actionable mutations and lower level WGS to capture structural rearrangements in cancers.


As for the deep sequencing of a gene panel, sure, why not, as long as you can get specific enough assays. As for low level WGS, to look for structural rearrangements...good luck with that.


Here's another article summarizing, in the abstract, what the point was above in my comment on how we can't surmise what we dont know yet.. In case anyone else decides its bullshiat, why not inform these authors.

Following the "finished," euchromatic, haploid human reference genome sequence, the rapid development of novel, faster, and cheaper sequencing technologies is making possible the era of personalized human genomics. Personal diploid human genome sequences have been generated, and each has contributed to our better understanding of variation in the human genome. We have consequently begun to appreciate the vastness of individual genetic variation from single nucleotide to structural variants. Translation of genome-scale variation into medically useful information is, however, in its infancy. This review summarizes the initial steps undertaken in clinical implementation of personal genome information, and describes the application of whole-genome and exome sequencing to identify the cause of genetic diseases and to suggest adjuvant therapies. Better analysis tools and a deeper understanding of the biology of our genome are necessary in order to decipher, interpret, and optimize clinical utility of what the variation in the human genome can teach us. Personal genome sequencing may eventually become an instrument of common medical practice, providing information that assists in the formulation of a differential diagnosis. We outline herein some of the remaining challenges.
 
2012-12-05 11:46:10 PM  

DO NOT WANT Poster Girl: Theaetetus: DO NOT WANT Poster Girl: The scope of the problem inherent in big data analysis on a single patient is enormous. We all have many millions of variations, including a proportion of copy number insertion and deletions. What particular variation causes what condition in a single patient? And don't even think that sequencing a single patients cancer tissue is going to get more useful info. There's too much info. How to cull the chaff? Nobody knows I don't know.

Good thing you're not everybody then, eh?

Sorry, but the biggest challenge in WGS is to make sense of several million variations in light of complex phenotypes. Are you someone actually working in this field ?


Tangentially, yes. I'm working for someone who's actually working in this field.
 
2012-12-06 08:08:05 AM  

Healthcare coverage
a short play.



The scene:
A mid-grade cube farm with consultation rooms lining the walls. Cheap office furniture. An extra with a bad haircut and an ill-fit uninform in the background watering plastic plants.

Insurance agent: "Hi there..."
pause as he looks down at papers
Insurance agent: "...Meagan."
Meagan: "Hi"
Insurance agent: "So you're looking to get some treatment for... um.. *flips papers* cancer."
Meagan: "Yes, I've recently been diagnosed."
Insurance agent: "I see."
Insurance agent: "It says here that you had whole genome sequencing when you were born."
Meagan: "It's possible. I don't remember."
Insurance agent: "Let's see. During that screening, it was determined that you had the proper genetic sequencing to cause this cancer, so technically it's a pre-existing condition."
Meagan: "Uh, what?"
Insurance agent: "Don't worry, we can't deny you coverage for that reason anymore, but our policies dictate that we have to move you to the high risk pool.Your rates are going up. Also since we knew it was only a matter of time before you got this, we're only going to cover the lesser of 40% of the costs of the treatment or the minimum the laws in this state allow."
Meagan: "Uh, what?"
Insurance agent: "Is there anything else I can help you with today?"
Meagan tears up.
Insurance agent: "No? Well, then have a nice day."

Que curtain.
 
2012-12-06 08:56:07 AM  
The ACMG guidelines recommends DNA microarray testing as a first tier test for postnatal analysis Link I certainly don't think that the preimplantation field is ready for WGA followed by array (due to problems with allel dropout and sampling error), but the postnatal world is certainly ready. As far as cancer panels, the accuracy is dependent on the technology (Illumina better than PGM) , sequencing depth and cutoff values (we don't call below 5%). If you have a panel that targets known actionable mutations (for either therapy or outcome) then this is an extremely reasonable approach. In fact, for many clinical trials, the sponsoring companies are going to require large panel sequencing prior to inclusion. FYI, I direct two clinical cancer genetics labs (cyto and molecular) and have skin in the game.
 
2012-12-06 11:35:54 AM  

moos: "Don't worry, we can't deny you coverage for that reason anymore, but our policies dictate that we have to move you to the high risk pool.


Which is also illegal.
 
2012-12-06 02:21:44 PM  

DO NOT WANT Poster Girl: As someone who does analysis of whole-genome data, both sequence and microarray,and writes statistical programs for analysis, I'm calling that expensive bullshiat.

There are no validated genetic tests I know of


I figured out your problem. The term you are searching for is called "pharmacogenomics." See for instance here Link Just like WGS, pharmacogenomics is in its infancy. This post is dead on:

Theaetetus: Keep living in the past, Subby. This is awesome for many reasons. Identifying particular genetic variants in your system leads to personalized medicine. Want a drug that cures your cancer without harming any of your normal cells? Have a mutation that results in inactivation of a protein that, on its own is fine, but means that you'll have horrible side effects if you take some drug that the rest of the population is fine with?
I have a client in this space, and they're doing some amazing things.


That's where we're going.
 
2012-12-06 03:09:18 PM  
thurstonxhowell:Which is also illegal.

It's not a documentary. It's a sarcastic joke about the difference between medical research and application (the other conversation going on in this thread,) the actuarial science/business combination that is Healthcare.

/but that's not important right now
//Do you like movies about gladiators?
///Don't call me Shirley.
 
2012-12-06 06:51:32 PM  

jengen: The ACMG guidelines recommends DNA microarray testing as a first tier test for postnatal analysis Link I certainly don't think that the preimplantation field is ready for WGA followed by array (due to problems with allel dropout and sampling error), but the postnatal world is certainly ready. As far as cancer panels, the accuracy is dependent on the technology (Illumina better than PGM) , sequencing depth and cutoff values (we don't call below 5%). If you have a panel that targets known actionable mutations (for either therapy or outcome) then this is an extremely reasonable approach. In fact, for many clinical trials, the sponsoring companies are going to require large panel sequencing prior to inclusion. FYI, I direct two clinical cancer genetics labs (cyto and molecular) and have skin in the game.


I just read that link and you do realize that their suggestion is to essentially SURVEY someone using a microarray to see if they can detect anything unusual? It's not to diagnose autism directly, or any other specific disease, only to see if they can detect something, and if something is found that's potentially interesting, they admit it's not always necessarily known to be clinically relevant. I won't repeat my post above, but it stands.
 
2012-12-06 07:04:46 PM  

lennavan: DO NOT WANT Poster Girl: As someone who does analysis of whole-genome data, both sequence and microarray,and writes statistical programs for analysis, I'm calling that expensive bullshiat.

There are no validated genetic tests I know of

I figured out your problem. The term you are searching for is called "pharmacogenomics." See for instance here Link Just like WGS, pharmacogenomics is in its infancy.


I've worked with pharma. I've been looking at marker-treatment associations for over a decade. The applications of pharmacogenomics right now are limited and largely tied into making clinical trials cheaper by trying to prescreen a target population, so fewer trials fail on non responders. Very conservative approaches with some limited research funding, and a feeling hat its the realm of academics to pursue. Notice that Illumina is partnering here.

The reason big pharma is not pouring more discovery money into pharmacogenomics is because there are too many variables to measure in terms of variations, and the costs and sample sizes are high, even in cancer research.

I would love it if there was more money available for research into genetics and genomics, as a bioinformatics professor type. The problem is that we are facing huge statistical variation in he genome cross patients, versus a limited patient sample size, plus biological noise in the assays we use to categorize our samples. Anyone who tells you different is blowing smoke up your netherworld.

That said, sure, we should sample more genomes, but without full IRB sanctioned patient data attached to it...better than the Framingham Heart Study ..not just tissue assay data, it's not useful in the long run.
 
2012-12-06 07:49:33 PM  

Theaetetus: DO NOT WANT Poster Girl: Theaetetus: DO NOT WANT Poster Girl: The scope of the problem inherent in big data analysis on a single patient is enormous. We all have many millions of variations, including a proportion of copy number insertion and deletions. What particular variation causes what condition in a single patient? And don't even think that sequencing a single patients cancer tissue is going to get more useful info. There's too much info. How to cull the chaff? Nobody knows I don't know.

Good thing you're not everybody then, eh?

Sorry, but the biggest challenge in WGS is to make sense of several million variations in light of complex phenotypes. Are you someone actually working in this field ?

Tangentially, yes. I'm working for someone who's actually working in this field.


Great, go ask them how they identify and /validate/ their targets. Most researchers targeting specific proteins do so with the help of information from animal models and/or targeted sequencing of human candidate genes p, often discovered originally in vitro, not whole genome sequencing or genomic microarray of single human subjects for a cost of 9K a person.

Amazing work is being done with WGS but sequencing a single person in interests of personal genomics is jumping the gun right now. There are plenty of pcr assays out there that can detect known risk variations...because they're known. And cheaper for anyone curious. More complex genotype patterns will require big..in the thousands...for patient sample sizes. The work to validate more complex phenotype associations in WGS and microarray should be taken by governments or industry, without charging a patient 9K for an nonvalidated diagnostic tool.
 
2012-12-06 08:18:26 PM  

jengen: The ACMG guidelines recommends DNA microarray testing as a first tier test for postnatal analysis Link I certainly don't think that the preimplantation field is ready for WGA followed by array (due to problems with allel dropout and sampling error), but the postnatal world is certainly ready. As far as cancer panels, the accuracy is dependent on the technology (Illumina better than PGM) , sequencing depth and cutoff values (we don't call below 5%). If you have a panel that targets known actionable mutations (for either therapy or outcome) then this is an extremely reasonable approach. In fact, for many clinical trials, the sponsoring companies are going to require large panel sequencing prior to inclusion. FYI, I direct two clinical cancer genetics labs (cyto and molecular) and have skin in the game.


Preimplantation genetic screening on targeted mutations in embryos using WGA was presented this year at ASRM. Presenters showed that the tech was at least equal to validated PCR assays. I was there.

I agree that for cancer a targeted cancer panel on known mutations would make great sense, but with WGS barcoding and the modest coverage depth needed if targeting limited regions....the presenters could detect with no problem with about 20x in the region....we're talking way under $9K a patient. And it's definitely better to target known mutations because a known panel is not a genome wide approach yielding many more unknown and non actionable variations.
 
2012-12-06 08:48:52 PM  
Sorry, I don't think I expressed myself well. As far as arrays for postnatal testing, the standard of care used to be conventional cytogentic studies, which only has a resolution of 5-10Mb. There were several studies published that showed CMA was the better first tier test, because there was a higher identification rate for abnormalities.

As far as autism, it is likely similar to MR/ID in that there are many ways to get to an end, due to so many genes being expressed in the brain. If you have a microdeletion you can have the phenotype of autism, but due to many different factors. By performing a CMA, you aren't looking for one thing, but just a high reolution virtual karyotype.

PGD is not my area of expertise and I was quoting what I remembered, in that for doing WGA followed by CMA there was potential for allele dropout and "overcalling" I think that looking for things like SMA or other deletions/point mutations has been done, but the recommendation is to follow up with an amnio. As I said, there are likely manychanges since I looked at this literature, so I defer to those doing this in clinical practice.
 
2012-12-07 11:03:07 AM  

DO NOT WANT Poster Girl: Theaetetus: DO NOT WANT Poster Girl: Theaetetus: DO NOT WANT Poster Girl: The scope of the problem inherent in big data analysis on a single patient is enormous. We all have many millions of variations, including a proportion of copy number insertion and deletions. What particular variation causes what condition in a single patient? And don't even think that sequencing a single patients cancer tissue is going to get more useful info. There's too much info. How to cull the chaff? Nobody knows I don't know.

Good thing you're not everybody then, eh?

Sorry, but the biggest challenge in WGS is to make sense of several million variations in light of complex phenotypes. Are you someone actually working in this field ?

Tangentially, yes. I'm working for someone who's actually working in this field.

Great, go ask them how they identify and /validate/ their targets. Most researchers targeting specific proteins do so with the help of information from animal models and/or targeted sequencing of human candidate genes p, often discovered originally in vitro, not whole genome sequencing or genomic microarray of single human subjects for a cost of 9K a person.


Maybe you misunderstand... I'm not working with someone doing targeted drug research, I'm working with someone doing bioinformatics. The "culling the chaff" problem you mentioned.
 
2012-12-07 02:17:40 PM  

DO NOT WANT Poster Girl: in vitro, not whole genome sequencing or genomic microarray of single human subjects for a cost of 9K a person.


I don't think we disagree on the big picture. I think you see what was written in this article very differently than I do and perhaps others as well.

DO NOT WANT Poster Girl: There are plenty of pcr assays out there that can detect known risk variations...because they're known.


Right. If you have brain cancer, doctors have some choices. They can try the targeted drugs one by one. Bouncing from drug to drug would be inefficient, expensive, silly and a pain in the ass. They can instead order one of the PCR based assays to look for specific abnormalities in order to more efficiently choose a drug. They could go from targeted test to targeted test until they find something. Or they can go genome wide from the get-go. I agree with you, the targeted tests seem best. But from the article:

"The price is dropping so quickly that it will soon become more cost-effective to do whole genome sequencing rather than targeted tests."

Right now it gives another option. You can go search test by test, or you can just do the whole thing at once. It's not so terrible. No one is suggesting a system wide change to WGS tomorrow.

You're also missing out on a huge goal of pharmacogenetics, that was brought up by Theaetetus.

Theaetetus: Have a mutation that results in inactivation of a protein that, on its own is fine, but means that you'll have horrible side effects if you take some drug that the rest of the population is fine with?


Some people take a drug and are "fine" other people are not "fine." The reason comes from your genetics. Do we fully understand the genetics behind the side effects? No, we hardly have a grasp on it at all. But correlations and statistics exist for specific drugs. We cant WGS you and say "Drug X will do exactly X, Y and Z to you." But we can WGS you and say "You're in a high risk population for Drug X leading to X, Y or Z in you so let's keep an eye out for that, k?" It's also possible we'll WGS you and say "yeah we tried and really have no idea still what will happen."

So should it be mandatory? No. But should it be optional and available? I don't see why not.
 
Displayed 27 of 27 comments

View Voting Results: Smartest and Funniest

This thread is archived, and closed to new comments.

Continue Farking
Submit a Link »
On Twitter






In Other Media


  1. Links are submitted by members of the Fark community.

  2. When community members submit a link, they also write a custom headline for the story.

  3. Other Farkers comment on the links. This is the number of comments. Click here to read them.

  4. Click here to submit a link.

Report