The paper uses Illumina reads. The "laptop" in the article is referring only to the platform for analysis, not sequencing.
4,759 karma · joined March 29, 2010
Opinions posted here are my strictly my own, and not representative of any employer or organization that I am a part of.
https://tedpak.com
[ my public key: https://keybase.io/powerpak; my proof: https://keybase.io/powerpak/sigs/fEW8iI_3tD3E9XjbSaPJTDOtaai7TXU6j1QAezE82Xc ]
Why? Three quick reasons: (1) Sequencing bacteria is really cheap since their genomes are 1000x smaller than a human's. Viruses are even smaller. (2) Bacterial genetics is much simpler than human genetics, and we have reliable, well-developed tools for testing hypotheses in vivo (i.e. you can grow many generations of transformed bacteria in a Petri dish while exquisitely controlling their environment, but human genetic manipulation is out of bounds for almost all scientists for ethical reasons.) (3) Results from sequencing pathogens are generally more actionable--you can detect transmission or drug resistance which can inform changes in patient care, as opposed to say, finding a putative increased relative risk of diabetes or heart failure, which doesn't translate easily into a plan of care.
I wrote a open access article about this here: http://cid.oxfordjournals.org/content/61/11/1695.long
Also see https://en.wikipedia.org/wiki/Public_switched_telephone_netw... regarding the 64kbit/s limitation.
VoIP, FaceTime, Skype and other IP-only technologies can use much better audio codecs, going all the way up to CD quality audio (44kHz, 16bit). This makes a huge difference when there is background noise and clipping and less than optimal transmission conditions.
(Source: I used to work for an IVR company.)
http://www.ncbi.nlm.nih.gov/pubmed/18344982
With regard to your second question, there are many "omics" assays that can now capture e.g. every mRNA, every protein, etc. at a given timepoint but you have to kill the cells. Fluorescence microscopy can work for "movie-like" time resolution on a live cell but you have to label the cellular elements with stains or antibodies and you can only utilize so many of those simultaneously on a single sample (1-5 is typical).
To my knowledge, no commonly sold cars are exclusively drive-by-wire for the braking system. If you lose all control in a car, you can turn off the ignition and forceful application of the brakes allows you to stop, even if you lose the power assistance.
By contrast, the most common failure mode for a drone in the air is... it falls out of the sky. That's pretty much the long and the short of it. There is good reason that flying objects should be held to a higher standard of safety.
Incidentally, besides the abdomen and pelvis, you can safely CT just about any other part of a pregnant woman.
That's why if you Google image search for "3D ultrasound", you get mostly pictures of fetus faces and very few medical conditions.
So, I see this product mostly benefitting the 3rd world, unless they vastly improve image quality and lowered the technical hurdles compared to existing ultrasound equipment. (Which would be awesome, but is a hard problem.)
Disclaimer: have used the Vscan as a student, but am no means an expert in ultrasound.
That's actually what I figured this article was about, based on the title.
I highly recommend this hybrid approach of 10-20 canvases plus a sprinkling of HTML elements for anybody looking to maximize responsiveness in this type of UI. In our tests, SVG and HTML start stuttering in draggable layers around tens of thousands of elements (easier to hit than you'd think). We actually decided to do mouseover calculations ourselves, in a pixel map within a mousemove handler, since having separate elements to trigger the mouseover was too expensive.
Wait, what? On production, install the exact ruby you need from your favorite package manager. On your dev box, install any rubies you need through rbenv [1]. Put all your gem dependencies into a Gemfile [2]. On either end, bundle install [--deployment] and call it a day.
You've basically described the immune system!
The immune system is imperfect, but it's a lot more complicated than any man-made system so far. Trillions of cells, thousands of genes controlling each one, hypermutations creating billions of different antibodies, slipping into and out of each corner of the body. Basically, organized chaos; by some measures, more complicated than the brain.
I would welcome potential designs for an auxiliary system, but what we have already is hard to understand, much less supplant. Certainly the concept of retraining the immune system is appealing, and looking beyond the OP's microbead/reinfusion rigamarole, we've already been doing that for decades via vaccines.
First of all, this hypothesis is still highly contested in the literature. There are no doubt certain pathways (usually part of cell cycle regulation) that happen to be used by most cancers, but there are hundreds of genes involved in the cell cycle and we aren't even close to enumerating all the ways that it can malfunction. The fact that most cancers use similar pathways is no doubt reflective of the fact that they are the easiest pathways toward malignancy, not the only pathways.
Secondly, say we do take the top N common mechanisms and create a bulletproof inhibitor. Now your statement about convergent evolution is no longer true. Anything that kills cancer cells (excision, radiation, chemotherapy, something targeted) exerts selective pressure, and you will instead start seeing cancers that utilize the (N+1)th easiest pathway, (N+2)th, and so on. It's like trying to stop all the ways that a car could break down: possible up to a certain point, impossible in the long run.
I also contest the statement that "you are basically reverting its evolution to day 0". Germline predispositions for cancer are an important part of the disease and not selected against in a society where people have natural-born kids between 20 and 40. If there were a cure for cancer, it would probably involve genetically engineering humans that are extremely unlikely to get it (and also age very slowly, etc.). But then you are looking at a totally different kind of society.
> genetic transfer to different hosts via some infectious vector but that is probably highly improbable
This is not only probable, it is known (so far, 12% of human cancers). Viruses linked directly to cancer include HPV, EBV, HTLV, and polyomavirus [1]. A virus that causes cancer can be a very successful virus, depending on how long the infectious/replicative phase is compared to the symptomatic phase. Replicating infected host cells = more viral production = greater likelihood of survival and transmission.
There are some researchers that think that all cancer is caused by viruses that have yet to be identified. I don't think that's literally true, but we can find all kinds of remnants of ancient viruses in our genome, so in a way, they might be on to something.
It is a fundamental principle of evolution that things that replicate faster and survive better will prevail. We are made of coordinated, replicating cells. They happen to be in a mutual "truce" to act as one organism at the expense of individual cell survival. Most cells that break the truce do badly on their own and die out. However, given enough tries at the prize, some cells will find a way to beat the system and replicate at the expense of the other cells still serving the whole body.
Some people have genetic deficiencies that give potentially malignant cells a jump start on the process. Certain environmental exposures will increase the rate of errors and therefore the frequency of tries that cells will get to become cancerous. But the outcome will always be the same; as we prolong human life through better medical care the incidence of cancer can only rise.
It would be most accurate to say that cancer is an evolutionary phenomenon that occurs between cells of a multicellular organism and causes a heterogenous group of diseases.
There most certainly are, at the information density of DNA. DNA is about a thousand times more information dense than our best hard drives. At that level, a passing photon can change or break enough chemical bonds to alter the bits. A nearby high-energy molecule can do the same. This actually happens gazillions of times per day in our body and we have a few hundred enzymes to automatically correct errors and a few hundred more to detect when those can't fix the problem and shut down the cell. The fact that it works at all is jaw-droppingly amazing. We take copying bits on and off hard drives for granted, and they only work for a few years max--cells operate on an entirely different scale of information transfer and yet some multicellular organisms live for centuries.
Unfortunately, I don't think you could improve the error checking mechanisms in our cells without fundamentally redesigning a lot of how we work. It would be interesting to try to add more "parity" mechanisms besides the complementary base pairing, which is used by most of the repair enzymes. From an evolutionary standpoint, no species would ever naturally develop perfect DNA replication because it would halt diversification, which is needed to survive continually changing environmental threats.
I should also add that it's a misconception that every cell in our body has the exact same DNA. That's only a half-decent approximation. In reality, a lot of our cells accumulate trivial mutations and it's OK. Some cells even rely on editing their own DNA to perform their primary function: that includes the B and T cells [1] which the cancer in the OP arise from.
I detest articles that use the word "cure" and "cancer" in the title. They are without fail so breathless about the clichéd achievement that they are exploiting in the headline that they either fail to acknowledge the real magnitude of the problem or gloss over it with practiced sleaze (not sure which happened to the writers here).
We've had a "cure" for cancer since the times when barbers also performed surgery on the side (Egyptian civilization, at least), and that is surgical excision. It's still the best treatment we have for many cancers, e.g. melanoma and lung cancer, both of which are much more common than the cancer in the article. It doesn't work for every cancer, but for some cancers it works good enough, and even better when combined with chemotherapy (which ham-handedly slows replication in every cell in your body). The point of any cure is to remove the cancer cells from your body and for some cancers they fortunately stay within one excisable area.
Like the GP comment pointed out, this is essentially a much more targeted version of the same thing. Obviously it would be better for the patient to retain B cells, without which they will have no adaptive immunity--just like a patient with melanoma would prefer to keep the hunk of their nose that the surgeon had to remove. This is an incremental, very fancy and very expensive improvement on the same old strategy of killing the organ to save the body. It will never work on non-blood cancers (which are the vast majority of cancers, BTW).
People do not understand that cancer is a disease that combines the trickiest parts of fighting aging with the trickiest parts of fighting infectious disease. Cell replication is one of the hardest things your body has to do, and it does it several trillions of times per day in your body, essentially copying about 1 billion TB worth of data while automatically detecting and fixing every dangerous error that could possibly result. It is natural that this process will screw up catastrophically at some point--on an infinite timescale, assuming we fix all other health issues, everybody will still get cancer just as surely as they age every year. And once they do, you have a cell that your immune system has carefully trained for decades not to engage, invading and hogging every resource it can, with mutations that allow it to adapt to selective pressures, including any drugs you might throw at it. Essentially, it's an infectious parasite, except it looks 99% like your own cells to your immune system, and is already perfectly suited to your body's style of metabolism.
Let's keep in mind that this can arise from any cell in your body that replicates, and just about every organ system has a good number of those to replenish malfunctioning or old cells and fix injuries. The possibilities are staggering and so are the number of known cancers.
There is no "magic bullet" that cures every single cancer, just like there is no "magic antibiotic" that kills every infectious bacterium. Even the smartest, most generalizable ideas right now, like cancer vaccines, depend on your immune system to make the final push, and the immune system is just as fallible as any other organ system. At the point where somebody can make the claim that all cancer is cured, we will as a matter of course have gained control over every replication event that occurs in our body (trillions upon trillions of nanoscale events per day). That will be a truly remarkable feat, but is in no way within reach of any foreseeable technology.
When we cure cancer, we will have by necessity cured aging and all infectious disease will have become a relatively trivial problem. That should put the claim of "curing cancer" into perspective.
However, this doesn't fix the problem with placing and canceling lots of orders to sway the bucket toward the end of the 5-minute window. I suppose if some could do this faster at the end of the window then others, they would still retain an advantage.
(Imagine somebody placing many orders, monitoring how many are likely to be fulfilled advantageously, and then cancelling the disadvantageous ones as close to the end of the window as possible.)
That could perhaps be solved with fees for cancelling bids or asks.
When I worked in Canada, I used the no-ATM-fee agreement [1] between ScotiaBank and Bank of America to convert USD to CAD and transfer money between the accounts. This involved the rather odd procedure of going to a ScotiaBank ATM, withdrawing a lot of CAD off of the USD in the BofA account, and stuffing the cash right back into the ATM as a deposit on the ScotiaBank account. Any other procedure, including using a teller or trying an online wire transfer, would have been more expensive and inconvenient. The above had no per-transaction fees and only a 3% conversion fee added to the market rate, which is about average compared to credit cards. (Any of Western Union's "instant" options for this kind of transaction require about 9% in fees.)
There are some credit cards with no foreign transaction fees, which is what I would use today if I had to do it again. Using such a card to purchase items in the foreign country, and then paying off the balance with the home country bank, results in "free" currency conversion. That's what GP is referring to.
As a commenter below put it so well, all of these responses are "a symptom of how we think about these kinds of problems being badly broken..."
Furthermore, I disagree that something is necessarily better than nothing. Wrong or incomplete advice can be much worse than no advice. Telling somebody it sounds like they have the common cold while missing out on the possibility of tuberculosis because the doc can't do a chest XRay, a PPD (skin test), run cultures, or listen to the lungs, is downright dangerous to that person and the people that they come in contact with. If a doctor then recommends the wrong drug to somebody based on incomplete information, the long term outcome can reduce or end a life.
This is why (1) medicine is already so heavily regulated (2) malpractice is such a prevalent concern among doctors and (3) it would only confuse healthcare consumers to endorse a second tier of medical care where the advice they receive might be "less right" than that of the first tier.
Here are some reasonable parallels to the dilemma you bring up:
- Plenty of people in the US can't afford to buy a car. Is it surely better to let them all buy cheaper used cars from foreign countries with crappy brakes, no seatbelts and no airbags?
- Plenty of people can't afford to buy meat as often as they'd like. Is it surely better to let them buy cheaper meat from unknown sources which hasn't been USDA approved?
It seems incredibly irresponsible for a doctor to prescribe things based on an internet conversation without a physical exam. You can't auscultate (listen to with a stethoscope), visually inspect, run labs, do basic imaging, etc. over an anonymous internet forum, which are basic facilities that anybody should expect a doctor in the US to utilize. It is plainly evident, once you become involved in clinical encounters, that you discover things on exam that the patient didn't know about, forgot to tell you, or wouldn't be able to find themselves. These findings are often critical for diagnosis. Telling the patient to take a certain drug without that data is dangerous and irresponsible.
You could certainly offer general "advice", but this will never be a substitute for seeing a doctor. Perhaps interfaces with video and sound are able to up the bandwidth of internet medicine but currently there is still too wide of a gap between that and actually laying hands on the patient.
I was not surprised to find that this is exactly what the linked document of related policies by state medical boards states for New York (http://www.fsmb.org/pdf/InternetPrescribing-law&policylangua...):
Section 80.63 of the controlled substance regulations requires a practitioner to physically examine a patient prior to initially prescribing a controlled substance. Issuing a prescription for a controlled substance solely on the basis of a questionnaire or other medical history submitted to a practitioner over the Internet does not meet the requirement of a physical examination or establish a legitimate practitioner-patient relationship and is not a valid prescription.
I cannot find anything controversial about that. I am all for forums linking doctors to talk to more patients, even for payment, but keep the "MD" out of the name because this is not a true substitute for seeing a doctor.
https://www.youtube.com/watch?v=AuXBcbYvKU8
uses an animation at 0:08 clearly from the STAR experiment at Brookhaven National Laboratory http://www.bnl.gov/rhic/star.asp which is run by Department of Energy (federal) funding and has nothing at all to do with startup technology, or New York taxes for that matter. The facility does happen to be located on Long Island.
A high-energy nuclear physics experiment focused around a mile-wide particle accelerator is so opposite to what startups aim to accomplish (minimum viable products, small teams, bootstrapping, becoming profitable fast) that the reference is nothing but humorous to me.
Perhaps the implication is that there are very smart scientists in the NY area that one can poach off to work at startups as federal funding for high energy physics research declines. :-(
At the lowest zoom level, one pixel is 280,000 bp (roughly the length of one or two genes, including noncoding segments). You can zoom all the way into the individual base pairs (a, c, t, and g).
There is a track below the chromosome cartoon (cartogram) that shows you the genes if you pull down on its label.
Cops are supposed to cite drivers for license plate covers and I see them out in NY the time--how can we expect them to enforce the display of their own numbers?
The only part of the uniform that police can't easily modify is the badge, which is also the part that unambiguously proves that they are police when necessary. (Not that most citizens would know how to authenticate the badge of their local police, much less those of all the federal LEAs [1]).
[1]: http://en.wikipedia.org/wiki/Category:Federal_law_enforcemen...
Which is a shame, because this means that convenience methods that Ruby has, e.g. ary.first → ary[0], ary.compact → ary.reject{|x| x.nil? }, ary.map → ary.collect are pruned out of the stdlib and frowned on in contributed libraries. This chilling effect that descends from PEP20 is one of the worse aspects of Python.
They increase readability and should be encouraged. Even if ary.last is one more character, it uses less of my brain to read than ary[-1]. ary.map might be more readable if other code uses ary.reduce, while ary.collect is more readable if other code uses ary.inject, ary.detect, etc.
The OP gave a perfect example with this---in an event handler for a drag operation within an editor, I'd rather communicate that text is being .highlight()-ed, even if the underlying view methods are reversing the pixels. If I used .reverse(), it might confuse a coder into thinking the text itself is being reversed when I drag.
Perhaps if more Pythonistas consider this a "best practice," it will swing favor for amending the Zen. But I wouldn't bet on it.
Also, you're incorrect about help(). help(Foo.highlight) will provide the docstring for Foo.reverse if Foo.highlight = Foo.reverse.