I'm a computer scientist working in a Parkinson's lab that is very closely tied to the names in the article. My main duty is data analysis simply because there is so much information to process. If you want some more information regarding state of the art treatments, our website is http://neurosurgery.ucsf.edu/bankiewicz .
I also maintain our website, so if you see anything that needs fixing, drop me a line :)
I'm interested in learning more about bio-tech research, genetics and synthetic biology. I hope to finish my doctorate in CS soon, and after that I'm going to take some time off to travel and read. Any advice for a computer geek who wants to learn how to program biology for the good of mankind? What do you, or others who work in the industry, think would be the best way for me to bootstrap myself in the area?
My lab brought me on because they realized that they needed to use computers more effectively with the amount of data they generated. I came in with little bio knowledge, but they did a good job of filling me in very quickly.
I'd suggest getting involved in any lab doing any form of computer work. Lab meetings are a great place to find out what problems people are having, and most of the time you'll be able to use your skills to solve them. After taking on a few projects, people will start to realize what you can and can't do, and they'll start to give you a lot of related work.
This can't be overstated. The shear amount of data available for processing is huge. Google-scale. I'm not very familiar with GWAS studies (I'm more familiar with high-throughput sequencing), but these are O(N^m) scale problems (where m>=2).
Let's see, how do I put this gently; the data generated by high-throughput sequencing blows the living hell out of GWAS data. One exome requires reading 30-60x of what's read in a standard GWA. (Not to mention that most people get GWA data on their next-gen sequenced stuff anyways.) You're in the most data-intensive part of the field... congrats! As you probably know better than I, we're still in the age of targeted sequencing and exome sequencing. Whole genome is just beginning to dawn.
Where I work, filesystem I/O is the rate-limiting step for most of my (next-gen sequencing) experiments...
Yeah, we haven't even had to deal with data directly from the instrument yet. So far, we've been getting data from collaborators for analysis via terabyte usb drives (FedEx throughput can't be beat). For the actual analysis, we've found the same thing... disk IO is a limiting factor. Well, that and the 16GB human genome indexes in RAM. And we aren't on an Isilon system yet (probably won't be either).
However, we just got our own instrument, so this will definitely be an issue, but our University knows a thing or two about dealing with big data (http://kb.iu.edu/data/avvh.html).
I've only dealt with a few GWAS style datasets, and the next-gen stuff dwarfs the GWAS data in terms of size. But when looking for linkages between variations, we're still talking more time than the universe is old level of calculations for more than 3 combinations. Which is really scary, because like you said, all the genetics people are going to be using sequencing for most things from here on out, so its like you have complexity on top of complexity...
OK, now that I've thought about it, the easiest way around this probably is throwing money at hardware (more disks) or optimizing the processing. However, this is only in the case of a true disk I/O bottleneck. If you're optimizing correctly then the disks should be reading 8GB blocks directly into memory and the CPU should be spitting them right out again. At the very minimum you should be using an optimized file system with large pages enabled in your kernel.
I don't think I have enough low-level knowledge to answer that intelligently at the moment. What I can say is that we're hitting these problems despite being on Isilon drives. (I think that is orthogonal to your solution, but again am not all too familiar with the subject.)
Not really... these datasets start to saturate 10Gb network connections very easily, so it's a question of volume. With some instruments, you can generate 10-20 terabytes at a time.
Just DNA, on my end (human lipoprotein genetics). I know that the cancer folks have a lot of interest in RNA, however. What is your focus? (Edit: your profile pretty much explains it! )
"Systematic brain delivery", is your lab simply jacking up L-DOPA or L-tyrosine in the substantia nigra? Also by massive information is your lab doing machine learning directly on fMRI output or do you also run statistical analysis on all published Parkinson's research in general?
Would it be possible for you to run a meta-analysis on Parkinson's studies involving alpha-7 nicotinic receptor agonists?
Our lab's main focus is targeted delivery to the brain which can be used to treat a lot of neurodegenerative diseases. Specifically for Parkinson's, we infuse an AADC vector into the putamen to break down L-DOPA into Dopamine. Further down the road, we are looking at using GDNF to save dying neurons that convert L-DOPA into Dopamine. The specifics are here: http://neurosurgery.ucsf.edu/bankiewicz/parkinsons.html
Regarding massive information, I analyze a lot of the data we generate in order to optimize our delivery platform. Since there are few labs that do this kind of work, all the data I work with comes from us.
If Sergey Brin thinks the current medical research establishment stifles certain kinds of research avenues, fine. But let's not cast this as him "proposing to bypass centuries of scientific epistemology in favor of a more Googley kind of science". He's not the next Francis Bacon, and this isn't some deep insight into the nature of science.
I think he is not proposing anything substantially new. Collecting data before formulating your hypothesis is the norm, not the exception.
First you see an object fall and them you wonder why. The same applies to birds flying, wood afloat, fire, atmospheric pressure... we observe(collect data) first, we make assumptions and look for more data later.
BUT, never underestimate the power of a GooglePlex for gathering and analysing data(like generic one). Never before we had used it, and so we could see things we never did before.
You know what happened when a man without studies created a miscroscope. The "scientist" of the time said that he didn't knew Latin, so he could not be called scientist, and that the microscope will only work for seeing the same things we already know bigger.
Because combing through large amounts of data in search of a hypothesis is as old as the hills. (In fact, the dangers of generating false hypotheses were already detailed on HackerNews a few days ago.) Brin isn't proposing some new way of doing science, he (I think) is just suggesting that the current mentality/funding-structure/whatever is not set up to support this type of analysis optimally.
Because he's not suggesting an experiment. This method involves taking data already obtained and doing analyses on it, to look for something, it's really a different way to form a hypothesis. After he finds something this way it'll go "back" to the standard rigors of the scientific method.
The world would be a better place if more young billionaires would have incurable diseases.
On a serious note, I hope they can find a cure for him and millions of others who can do nothing about their illness other than playing a waiting game.
The world would be a better place if more young billionaires would have incurable diseases.
They do. Aging may not technically be a disease, but it's currently unavoidable and cripples your body and often your mind. If I had a few extra billion dollars lying around, that's what I'd focus on.
What's really interesting about this is that he doesn't actually have the disease. Due to genetic testing, he may have an increased probability of getting it.
>Many philanthropists have funded research into diseases they themselves have been diagnosed with. But Brin is likely the first who, based on a genetic test, began funding scientific research in the hope of escaping a disease in the first place.
I'm surprised that he didn't invest more than 0.3% of his net worth (if it really is $15B) into this. It's his choice, but I think that if I was in his position, I'd spend more than that (but yeah, I guess that's easier to say when it's not you).
Maybe there are bottlenecks that make it so more money wouldn't speed things up, but I bet there are ways to use resources to help alleviate or remove these bottlenecks.
Why would it have to be a percentage of his net worth? He's contributed $50MM which he feels is enough to "move the needle". Medical research moves slowly and simply throwing money at it may not offer any incremental benefit.
Edit: I thought I'd add an example, since my father has Parkinson's so I know a little bit about it. One promising area of research is stem cell therapy. However, one known side effect of stem cell injection is cancer. (It takes years to develop but there have been enough studies already to show a strong correlation). Now, Parkinson's doesn't kill per se - it makes life miserable towards the end, but as a rule, people with Parkinson's live just as long as people without. So let's say you had really high net worth and wanted to spend it all to cure Parkinson's. Would you be able to "buy" people to trade a low chance of improvement in their lifestyle for a very real chance of cancer? I don't think so. Of course, this is just one example, however, it shows that things are not that simple.
I'm just saying what I would do if I was worth 15 billion and knew that I had a high chance of having a serious disease like that.
Throwing money blindly at a problem doesn't always help, that's very true. But I'm sure there are many intelligent ways to use more than 50m effectively.
You mention stem cell therapies; helping get them to the point where they don't cause cancer anymore might be an area of research worth pushing, especially if some of it is under-funded. Giving out scholarships to people studying in areas related to the disease might be a way to help improve the supply of researchers. And when you can't hire more people, you can usually upgrade the equipment with faster/better stuff (shortening the upgrade cycle might help accelerate progress). You could help pay for conferences where the leading minds working on the problem can meet and exchange information (and if there's already enough of those, you can help make them better). You can probably find high-risk/high-reward research projects that aren't getting funded through the regular channels because those usually favor low-risk/low-reward projects. Etc.
But I repeat: What Brin did is awesome. Kudos to him. All I'm saying is maybe I'd do it differently if I was in his shoes.
Thank you so much for posting the video! (For those who might want to watch it also, the interesting stuff starts at the 5 and a half-minute mark). The re-framing of cancer as a desirable injury response is truly eye-opening.
The amount of money you give doesn't matter. What matters is who you give it to and for what purpose.
Specific conditions like Alzheimer's get massive research funding. But if the majority if that money had been invested in research targeted to expand basic knowledge of brain neurobiology, then we would be MUCH further along on real treatments that are actually effective.
When you invest heavily into research for one specific neurodegenerative disorder you are investing in specific and superficial symptom treatment. Investing in neurbio research expanding basic knowledge of the brain instead will give you more results in treating a wider range of neurodegenerative disorders.
In this specific context, I don't think appraising the total money Sergey Brin has given matters. I'm sure he has enough money to fund any/all decent research proposals. The decent proposals must come first though, one can't simply say he should be throwing X total number of dollars at a problem. Throwing more money at things does not always makes things better and sometimes can make things worse.
I would guess that a) spending rapidly reaches the point of diminishing returns and b) it's probably very, very hard to effectively target research money. As such, it might simply not be effective to spend more.
In addition, he'll probably continue investing in research going forward; $50M to date might eventually increase depending on what's found.
The article sites the average American's risk of developing Parkinson's at 1%. I really question that. My husband's father died of Parkinson's several years ago, my own father is currently living with Parkinson's, my best friend's father has Parkinson, and so on. Am I the only one surrounded by Parkinson's disease, or is it a growing epidemic? I wonder if that 1% number is an old statistic. I can't find any sources, links?
"Random is blotchy". At 1 out of 100, you don't have to know that many people who are able to get Parkinson's to have a non-trivial chance of seeing "more than expected" the number of cases. I don't know anybody with Parkinson's, and I "should", I know well more than 100 people who could have it, and if they did, I would know. And of all the people reading these comments and able to post, the odds that at least one would be in a "comment-worthy" blotch is basically 100%.
Doing the math (450,000 cases out of 36,000,000 million Medicare recipients in the US), the adult incidence rate of Parkinson's is 0.0125, or 1.25%.
Sadly, I can't find a full copy of the study which isn't behind a pay wall. But, from the summary, it appears Parkinson's has some "clustering" based on both genetic and environmental factors. The text under the map on the article indicates areas in red have an adult incidence rate of 13.8% or more, so experiences such as yours are normal (albeit frightening).
So, yes, you're right to question that 1% statistic. It isn't sufficient to describe the situation at all when standing by itself.
Thanks wooster. You are the only person who actually answered my question. I guess this forum has a chance :)
I wonder if with the "graying" of baby boomers, the statistic is going to shoot up. I wouldn't be surprised if by 2020 the numbers would double, along with the projected retirement numbers. Of course, that would mean that the cluster areas would see rates close to 1 in 4 older adults, which would be truly scary!
Well, as a competing anecdote, I don't know anyone who has Parkinson's. And I haven't heard anyone I talk to in person say they know someone with Parkinson's.
You are confusing your personal experience with statistics. You can trivially find sources. Every single person you know might (unfortunately) have Parkinson's while the US population incidence remains around 1%.
You are confusing your personal experience with statistics.
We are not a bunch of 5-year olds around here, and I in particular am not so primitive as to think that the world revolves around me. There is nothing wrong with questioning mainstream journalism, especially when it's without citation and contradicts personal experience.
You can trivially find sources.
If so, wouldn't it be easy to paste a link, which is what I've requested in the first place? I hate the knee-jerk "google it" response - the person who gives it usually thinks that they are smarter than the person who asked for help in finding something, but all it shows is that you are (a) lazy and (b) disrespectful.
Now, if you can provide sources, please do, and if you cannot, I respectfully ask that you please leave this thread to people who care to contribute something other than "google it, stupid".
I am curious, are you being sarcastic? You just asked me why I would get upset when someone "deigns" (dict: to behave as if one is descending from a superior position or rank) to "correct" me (dict: to rebuke or punish in order to improve).
Your choice of words aside, I care about this forum, and I don't want it to become the usual deal where people assert their ego at the expense of actual discussion. My intent was not confrontation, it was cooperation.
Edit: Never mind. I see from your "ha ha ha" post below that you are trolling. I guess that answers my question.
Good link. At the end of the day, they're doing association studies... So they're taking a phenotype (even survey answers are a "phenotype" in a manner of speaking) and correlating it with a genotype (whether SNPs, exome data, or genome data). I'm a believer in this approach to science and am in no way poo-pooing their effort; I merely wanted to comment that, from the article's description of their approach, they're doing nothing iconoclastic.
I think if I had that kind of knowledge of a fucked up disease hanging over my head, I'd put more than .3% of my net worth into the research. I'd want to REALLY move the needle, like alot.
IRB review is a very important part of modern science, and the system is simply not setup to deal with research that hasn't been throughly cleared by a review board. You could consider IRB approval the first step in peer-review: your study is worthwhile and you aren't going to do anything unethical.
IRBs are designed to protect the subjects/patients, the researchers, and the institutions. To skip that step is pretty unusual, and some hoops had to be jumped through in order to deal with it. I think the fact that the 23andMe study was published at all was a big leap, and represents the flexibility inherent in the scientific community.
The response to "new and cool" wasn't "you didn't do it according to protocol". It was: since you didn't do it according to protocol, we need to verify that you did this ethically.
Imagine what would have happened if they used someone's DNA in their study without their knowledge. When you're talking about publishing links between diseases and people's DNA, you better be damn sure you have permission to use their genetic data. That is what IRBs are good for.
In a world where movies constantly are pointing out how scary/evil/unethical scientists can be (Andromeda strain, Splice), these checks are very important, even if it slows things down a little.
> Imagine what would have happened if they used someone's DNA in their study without their knowledge. When you're talking about publishing links between diseases and people's DNA, you better be damn sure you have permission to use their genetic data.
Apart from privacy and "ownership of one's genetic material", why?
Saying "Joe Smith's DNA contains a vital clue to solving parkinson's" doesn't affect Joe's ability to monetize.
Saying the same about "unspecifed_37" doesn't affect anyone's privacy.
There was a good episode of Radiolab about her, as well as a good interview on Dr Kiki's Science Hour with the author of the book written about her life and legacy.
Her family was deeply disturbed by what was happening at first. Progress is necessary, but so is people's dignity.
Her family was deeply disturbed by what was happening at first. Progress is necessary, but so is people's dignity.
I don't want to be callous, but look at the cost-benefit calculation:
Cost: One disgruntled family.
Benefit: Polio vaccine. Treatments for cancer, HIV, various toxins, and other stuff I don't know about in the 50,000 research papers on HeLa. I am not exaggerating. There really are over 50,000 papers about HeLa.
That doesn't really change the fact that there were issues with her informed consent. When HeLa cells were first derived, most of the benefits weren't even on the radar. I don't think the initial investigators could have imagined what would have happened, let alone the patient.
I think this is a case where the system didn't work... but more because the science got ahead of the ethics. Sure, excellent science has been done with HeLa cells, but it all started with an ethically ambiguous beginning. For example: why do we know her name? This alone should never have happened.
When HeLa cells were first derived, most of the benefits weren't even on the radar.
Excuse me? It was the first human cell line that could grow in culture indefinitely. Doctors and researchers immediately recognized the potential of HeLa cells. It took less than three years before the cells were used to mass-produce Jonas Salk's polio vaccine.
I am not exaggerating when I say that millions of people today owe their lives to HeLa. Had doctors been forced to do what you call ethical, those people would be dead. With so many lives at stake, the objections of the family (or the individual) simply don't matter.
For example: why do we know her name? This alone should never have happened.
Originally she was anonymized. Later, researchers wanted to get genetic information from her living relatives. Somehow the press found out. Perhaps a relative told them.
The thing here is that this is the start of a different way of doing science, and as such you want to be careful of the precedents you set.
It's a good thing that someone is thinking about how all this stuff is going to be protected.
The protocol there is one of ensuring anonymity for people that want to remain anonomous. Making sure that that sentiment is retained moving forward is not a bad thing.
I do not read the editorial you linked as saying, "You didn't do it according to protocol." The editorial, in fact, is explaining why PLoS Genetics thought the study was okay to publish, how the 23andMe study strictly complied with existing protocol and went above and beyond its requirements, and how the IRB system needs to be improved to deal with GWAS.
"Cigarette smokers also seem to have a lower chance of developing Parkinson’s"
I realize it's not really a science article (it's even possible all the information on Parkinson's came directly from Brin himself), but they didn't say cigarette smoking staved off Parkinson's completely - they said it lowered the odds of it developing.
Thing is, the human body is so complicated, you can do all the right things and still come up on the short end. Load up on omega-3s and still get clogged arteries. Eat a well balanced diet and still get cancer.
Doing the "right" things can only improve the odds - something's going to get to you in the end.
"Thing is, the human body is so complicated, you can do all the right things and still come up on the short end"
It's complicated, but it's not magic. If you are doing things to avoid a problem and you get the problem, then you were not doing "all the right things" - your understanding was wrong.
> If you are doing things to avoid a problem and you get the problem, then you were not doing "all the right things" - your understanding was wrong.
Nope. "All the right things" isn't a guarantee and since you can't do anything about your dna, some folks are going to have risks no matter what they do.
For example, you can get unlucky wrt a mutation or a cosmic ray.
I also maintain our website, so if you see anything that needs fixing, drop me a line :)