apathy

> confirmable

But were they indeed confirmed? This is the problem that keeps cropping up. Sometimes I wonder if believing in findings from the psychological literature isn't a disadvantage. Knowing about them is good -- you can pick up on the chumps who actually believe them. But believing them?

It's not at all clear to me that uncritical acceptance of these studies is ever a good idea. The response of senior investigators shrilly decrying "methodological terrorism" when their small-sample findings fail to replicate does little to increase confidence in said findings (or, indeed, the field as a whole).

nb. In no way is this confined to psychology. It's just more rampant there, to judge by the foundational manner in which the field has been shaken up. Nobody doubts that PCR works or that iPSCs can be induced, because the experiments to replicate the results are relatively trivial. But the nature of psychological research seems to make such exercises very difficult, to the point that individuals "take it on faith" that an effect is real (as opposed to assuming it's wrong and running a quick experiment to show that the next year or two of their life won't be wasted). Maybe we're just more cynical in bench-based disciplines?

The highest-marginal-return skill to practice, in my experience, is to figure out where the mutual benefit in any arrangement is, and PRESENT THAT FIRST to each party involved. If it seems like things magically improve whenever you show up, most of the other issues go away.

Also, ensure that whatever you do makes your "superiors" look better, but make goddamned sure that it reflects their cleverness in choosing YOU for the job. If you're in academia and you write 99% of a grant that gets funded, put it on your CV. If you write a contract or a proposal that makes the company $5m, put it on your resume. Don't ever expect anyone to give you credit -- take it, and then thank whoever would logically be obligated to give it for their gracious acknowledgement. People seldom do what they know to be right -- they do what is convenient, then repent. So make it convenient for people to do the right thing for you.

Make it easy for people to appreciate your value and hard to fault your shortcomings. It goes a long, long way. If you can't do that, maybe look for a position where it will be easier for you to execute this task.

Good luck! You will be a happier person if you do this.

That's not how tenure works. Up or out -- you get tenure or you hit the road. So this is actually fairly common in academia. Sort of a parting shot, if you will.

Presumably about n-fold higher (marginally) than without redundancy. Not counting the cost of this sort of outage, which could swing the equation strongly in favor of nX

other professions do behave like this. Look at the CEO of Wells Fargo or Valeant. "Fuck you, I got mine!"

if Google is down you'll have other problems (SYN, ACK, SYNACK won't help much with the DDoS that's killing it)

> If you assume power of 80% (not crazy in rigorous studies) and 50% of programs having an effect...

You're making another assumption: that the power calculations are actually rigorous and the effect size (not "having an effect", but "having an effect of at least the size used to calculate power at a given alpha") is sufficient. If you've ever designed a clinical trial or written up the statistics for a grant, you know damned well that these numbers are cooked 6 ways from Sunday.

Colquhoun is a sharp guy (I have debated various fine points with him on several occasions and he has convinced me with reproducible examples that I was wrong in my beliefs). It would be nice if the name of the article was "It's time for science to abandon the term statistically significant", as in the URL, because that's the real point here.

"Signficance" can only be judged in context. Post all your data and then we'll see whether you can be believed. But that notion of transparency is absolutely terrifying to the sorts of senior scientists who control most funding and peer review at the moment. It's OK, though; the rest of us put up preprints knocking down the most egregious lies and those who care about the truth can have it (for free, no paywall).

I've got plenty of Cell, Nature, and NEJM papers on my CV; direct support on grants where I am PI, Co-I, or KP is over $5M; and I still think the situation is fucked up. I'm not bitter because I feel left out; I'm bitter because I fear that the bad money is driving out the good, as it always does.

Take away the monetary incentive, the Journal Impact Factor BS, the "prestige" of hoodwinking 3-5 referees with a pile of STAP or single-sample comparisons (had to warn a student about this when he pulled some data from a Nature paper that a respectable MSKCC scientist gooned for...) and let's see what's left...

nicely done, hope it works out as a business for youse guise

I see. Where's your rebuttal regarding nonavalent Gardasil, the actual topic of conversation here? When you can provide some evidence (any evidence, really) that the immediate risks outweigh the benefits, I'll be impressed.

Meanwhile, enjoying the benefits of herd immunity while other people and their kids take the risks is the definition of cynicism.

I really ought to clarify this: yes, we absolutely should do follow up studies to make sure HPV vaccines are doing what we think they are (Gardasil for example).

I do not under any circumstances think it's appropriate to delay HPV vaccination incentives to do these studies. If you think about it, the only perfect model organism for humans is humans. As long as it's ethical (IMHO it absolutely is), the right choice is to proceed and monitor; doing both will provide the appropriate data to judge whether we are accomplishing the goal.

Genius. Post a bond.

The only reason you know that Tamiflu is useless is "cynical monsters" like me who bite the hand that feeds them.

As far as spacing doses, just like anything else it's subject to revision.

Nonavalent HPV vaccines produce antibodies against nearly every HPV variant known, reduce the likelihood of HPV integration, and for someone who has verified that at least a quarter of head & neck squamous, almost half of anal, and far better than half of cervical carcinomas show HPV integrations, typically right into the middle of tumor suppressors or onto oncogene promoters, this is a big deal. These cases are preventable. Not preventing them, when it's relatively cheap and available, is (to me) APPALLING.

The antivaxxer pricks are cynical for relying on the rest of us, who are responsible enough to assume a TINY risk to our kids in order to protect them and others, to take the risks so they can enjoy the benefits of herd immunity.

If that isn't the definition of cynical, I don't know what is. It's on a level with the Soviets celebrating the "eradication" of smallpox while internally working to weaponize it.

Provided a child develops appropriate immunity by the time they're regularly interacting with others, I could care less about the spacing. But that's not really what this is about, is it?

If you have kids and you don't get them vaccinated on an appropriate schedule, you are being a cynical prick. If you fight HPV vaccination because it might "make girls promiscuous", you are a cynical prick. You are presuming to judge risks for others, risks which we as a society have determined are unacceptable. And that makes you a cynical prick, if indeed you do it. (Maybe you're playing devils advocate, I don't know. But I really hate this line of reasoning, and on balance, it doesn't hold water.)

Hey if they have their way, we'll all "continue to study" HPV vaccines so that more girls can die of cervical cancer.

Fuck those cynical monsters. I don't wish people ill lightly, but they're taking chances with the lives of their children and of others, and I just can't stand that.

Hey, just so I'm not a total hypocrite, it turns out that you can "turn off" PD-1 and PD-L1 response in CAR-T cells (at least in mouse models, https://www.jci.org/articles/view/83092 ). In the end I stand by my contention that it's a tool and it needs to be used for the right jobs, but it's a very versatile tool.

I work on immunosenescence and immune stimulation (the real thing) in oncology, and I happen to believe that we do many patients a disservice by nuking their immune systems with chemo and then relying on things like TKIs or BiTEs to pick up the mess. But it's also important to keep in mind that Gleevec alone didn't get rid of CML, and the way ALL and APL were managed up to 95% cure rates is by adding a little poison (As2O3, methotrexate, etc.) to the brew (asterisk) to finish things off. It's a delicate balance. I think it always will be.

(asterisk: ok more like a metric fuckload of poisons for ALL. But the real trick is "resurrecting" the patient with leukovorin after folate starvation essentially kills them.)

http://blogs.sciencemag.org/pipeline/archives/2016/07/08/car...

Look up the SuperMAB saga for more cautionary tales. Plenty of people have died from chemo and immunosuppression too (nobody benefits from lungs full of aspergillus and a gut full of C. diff). If it were easy, somebody would have figured it out already :-)

Spoken like someone who hasn't been following the Kite trials...

Also of note: just like good ol' mAbs, you need a target for CAR-T cells. And BiTEs, and BiKEs, and what have you.

Also also of note: the approaches that took pediatric ALL and adult APL (among others) from 5-10% survival to 90-95% survival weren't single pronged. (Good thing, too; I've seen patients who seemed ideal for CAR-T die when their malignancy outran the engineered cells) The majority of progress in real-world outcomes has arisen from combination or sequential therapies, incrementally improved to decrease toxicity and maximize efficacy.

Which is to say, incrementally advanced.

Last: per the above, you can use the two together. If a tumor is actively suppressing T cells, it isn't going to dissolve from CAR-T cells without a fight. There are additional tactics useful for "priming" the immune system, but in the end I don't see a lot of solid tumors melting away in trials of CAR-T cells (feel free to correct me).

Do take a look at the Kite trials and how they're structured. When you have to condition a patient with fludarabine to hit outcomes (at all) that's not a wonder drug. It's another tool in the toolbox. A necessary and useful tool, but a tool, not a magic wand.

The majority of pediatric malignancies are either germline related, in utero (de novo mutation/SV, potentially caused or facilitated by maternal environmental exposures), or a reverse lottery winner. There simply isn't enough time for somatic mutation to cause the sort of devastating fallout that you see in DIPG or infant leukemias. (Furthermore, even the point mutations seen in pediatric cases are characteristic and rare or absent in adults; some structural variants are also observed in adults, but they are much rarer and accompanied by fewer cooperating events)

Armitage-Doll was a nice advance over previous models, but it's also incomplete (and probably flat wrong in some cases, although working on pediatric malignancies has convinced me that a second cooperating event usually is mandatory).

In normal stem cells, it appears that attrition and immune clearance gets rid of damaged cells when they cycle and senescent cells all the time (subject to some variation, not entirely age related, at least in our volunteers). We may expect higher rates in filter organs, but liver cancer isn't too common, and the paper I referenced earlier shows that this can't be just an issue of fewer divisions (I despise the oversimplified Tomasetti & Vogelstein paper because the facts simply don't support it). Colorectal is probably more common because the crypts are "facing out" ala melanocytes, thus prone to accumulating lots of environmental damage.

Anyways, the latter of your possibilities (proliferative mutants divide faster and error more often than normal counterparts) makes the most sense -- the eventual "winner" in a tumor is the cell that produces the most progeny and resists apoptosis due to stress the best. It's probably not a coincidence that these are traits which adapt a mutated cell to survive chemotherapy as well. However, spawning nonself mutations willy-nilly is a great way to attract immune attention -- particularly if you haven't blown the immune system away by nuking it with chemotherapy. :-/

Write error == (fixed) base pair mutation. A classic example is a methylated cytosine spontaneously deaminating to yield thymine. Our DNA repair enzymes can't tell the difference (in terms of which is the "right" base) between the thymine and the guanine left behind, but since they don't match, one of them has got to go. Thus there is a 50-50 chance that the mutation will be fixed. That's the easiest example because it's not an "error" per se (rather a thermodynamics problem) but genuine proofreading errors also occur. The net rate is about one in a billion bases.

These estimates ignore indels and SVs but empirical evidence suggests that the "everyone over 50 has a 50/50 chance of at least one adult stem cell harboring a mutation in at least one interesting gene". My personal favorite is

http://www.nature.com/nature/journal/v518/n7540/abs/nature13...

but Druley's follow up was equally awesome:

http://www.nature.com/articles/ncomms12484

and the recent survey of adult stem cell mutations is nice:

http://www.nature.com/nature/journal/vaop/ncurrent/full/natu...

One take-away from all this is that, while 95% of a sensitively surveyed population of 50-60 year olds had at least one stem cell with at least one known preleukemic mutation, it is equally clear that most people aren't walking around with anything resembling an acute leukemia. The natural conclusion is that in individuals with a competent immune system and diverse enough pools of healthy stem cells, it's not that big of an issue. Only when bad luck and/or stresses to which the mutants are adapted (e.g. TP53 mutations in therapy-related leukemia) afflict people, or the natural diversity of their stem cell populations collapses (as with really old people and individuals whose immune system actively attacks their stem cells, as in severe aplastic anemia) do you see the sort of massive, life-endangering takeover that we recognize clinically as disease.

Furthermore, nearly all of us are born with 5-10 predicted-to-be-lethal variants in our genomes. Clearly, we're also not dead, so our conception of "lethal" can't be quite right. There is an enormous amount of complexity in how real live multicellular organisms deal with variation and mutation, something we're really only just starting to grasp, and of course all of that then interacts with the person's environment to manifest (or not) their genetic tendencies. We build models of reality because the actual thing is too complicated to be tractable; it's important never to confuse the two :-)

It's not 100% clear to me that this is directly compatible with the graph reference, but if it isn't, that seems like a simple matter for Batzoglou & co. Those guys are really, really good, and the multi-reference target is a type of graph anyways, so I imagine it's just a bit more generalization (if any) to make BWBBLE work on a graph-structured reference assembly as implemented (there seem to be certain inconsistencies about how to represent structural variants that differ between implementations).

The other thing that would be neat is that then you'd have a direct tie-in to ancestral recombination graphs and could, in principle, get IBS/IBD for the same cost as high-confidence genotyping for any two individuals. Come to think of it, there's probably a way to recast this as shortest paths and get all admissible traversals between a population of genotyped individuals (given an ARG) for the same price as any two. Hmmm. This is a little disturbing.

There's probably, in spirit, a graph version of the BWT, but I'm not familiar enough to know it. The approach is less straightforward because you're not just modifying BWT to allow for indels/errors, but rather a graph compression algorithm to allow for fragment searching along vertices of a sequence graph. That said, it has to happen eventually. It's been a while since I took comp bio (sequences & graphs) but what you mentioned is what I learned (and I took it from Waterman, so what I'm saying is that I think you've got it).

You can look at Pall's work to see how the current approaches may evolve into something compressible:

https://github.com/GFA-spec/GFA-spec

As far as articles? McVean's, without a doubt!

http://www.nature.com/ng/journal/v47/n6/full/ng.3257.html

An implementation paper for graph assembly HLA typing is at:

http://biorxiv.org/content/early/2015/12/24/035253

Interesting times ahead, with people recognizing that GxE and GxExE matters far more than G alone.

Google is using graph learning for representations of stuff where obtaining a huge, representative, "labeled" corpus is hard.

As many in the field can tell you, "labeling" genomes is hard, too. You can stick labels like "died at 39 of AVCR" or "had 5 primary tumors by age 28" or "insane multi-substance abuser with extra toes" on a person, but that doesn't really encapsulate all their traits. I claim that genome analysis is a great candidate for semi-supervised learning. Looks like Ben Paten already had that thought...

Haussler, Paten, McVean, and the usual suspects are working on a tractable graph representation to replace (say) hg39, i.e. instead of a new "reference genome", there should be thousands. This makes more sense when you look at how common structural variants (say, inversions) are, and then when you do something like ask Haussler at ASHG "how do we represent inversions in this thing?" you realize how fcking hard it really is to get it right.

McVean & co. made it work for the major histocompatibility complex, which immunologists can explain better than I can, but that's perhaps the most diverse bit of genome there is. It's riddled with ancient repeat element insertions and generally fascinating. It's also a source of Little Problems like organ transplant rejections, graft vs. host disease in stem cell transplants, maybe bits of schizophrenia and neurodegenerative disease... in short, if it can be done for the MHC, it's probably doable for the whole enchilada. It's not easy, though. In fact it's so difficult to get right that a sub-field of accurate HLA typing and immunotyping evolved in parallel to large-scale genome analysis, because it's really, really important not to fuck this up. So that's a note of caution from historical precedent.

Nonetheless, it turns out that your pals at Google spearheaded the effort to have a Hadoop-like "data lake" for genomes (sign up for the GA4GH mailing lists if you like to watch professors bikeshed an API, and occasionally produce incredible insights by accident). Maybe this is going to converge in an interesting way. It won't happen overnight, but it will happen, and the mathematicians will be vindicated.

* Seven Bridges is a little genomics company with a funny name. Unless you're familiar with Eulerian paths and the Seven Bridges of Konigsberg, in which case it's sort of obvious why they chose their name. Nice people, other than the patent, which infuriated most everyone else.

Proofs are math, science is uncertain (otherwise why bother running experiments)

That's long for math

The best part is that any fool can see that impact factors are a terrible proxy for real impact, and in fact the editors of several glam journals have pointed this out (disclosure: I have plenty of glam papers on my CV*).

But since it props up a myth that senior faculty like to believe (i.e. their shitty old Cell paper is great because Cell is/was great) that's the yardstick. As the director of a CCC once told me, it takes too long to see the impact of papers (citations piling up) so the JIF is used as a proxy.

Sort of like how it takes too long to do good science, so some people just publish whatever garbage they can sneak past the editors (ha fucking ha only serious). There is a LOT of crap in the literature as a result.

In general, if you have to put science in the name, it isn't.

neither are bullshit papers/editorials/letters-to-the-editor

http://frog.gatech.edu/Pubs/How-to-Publish-a-Scientific-Comm...

"Eminent" specialists often argue in favor of eminence (journals, tenured investigators, blah blah) over evidence (tossing an analysis up on arXiv and posting a link to it). Young turks or iconoclasts often do the reverse, since there is a power structure in place via editors, reviewers, study sections, and the like. Orthodoxy often sticks even when the evidence supporting it is quite skimpy (or absent).

Funny thing is, when you pull the underlying data (in fields where this is possible), you routinely find that the reported conclusions are overblown. Not necessarily wrong, but routinely sold as more conclusive than they are (and sometimes they are in fact wrong, whether due to small sample sizes, "outlier removal", or outright fraud).

In no way, shape, or form are these bad habits limited to psychology. They happen in basic biology, they happen in medicine, they happen in clinical trial reporting. It's faster and easier to oversell shitty science than it is to do good, thorough science, so with the tight competition for grant funding, you can imagine what happens next.

Trust whoever you like, but verify results from everyone. There are good people out there who are fastidious; there are good people out there who are sloppy; and there are people who don't care one bit whether they're publishing absolute horse shit. The onus is on you, the reader or researcher, to do the requisite critical thinking (and, perhaps, a few analyses before you waste time running down a dead end).

Keep an open mind, but don't let your brain fall out. Also, many good scientists aren't very nice people, and some very pleasant people in science are shitty scientists. It's very hard to tell a priori which is which, so look at the evidence and decide for yourself.

Counter example: yamanaka

Specific only in that the categories aren't supervised.

Furthermore, suppose you have labels for some but not all points on your data (i.e. your model is designed to be robust in the face of things it hasn't been trained for). There are a nontrivial number of people who work on either side of the "semi-supervised" divide, e.g. clustering with examplars or pulling out the generative model for a discriminative task. Personally I like these better, as they're more akin to what people seem to actually do (encounter new things and try to make sense of them).

Anyways. If you look at the delta in performance between "old" techniques like random forests or gradient boosting vs. deep convolutional networks, it tends to be quite small until your datasets grow to very large sizes. For things like images that's not much of a problem. For things like rare diseases it's a huge problem.

I never would have tried MonetDb (ok monetdblite) if not for this great little tutorial on how to load all of SEER into it:

http://www.asdfree.com/2013/07/analyze-surveillance-epidemio...

Yeah the presentation and code isn't beautiful, but it does avoid the need to WRITE THE DAMNED THING YOURSELF, which some people apparently will never understand (although they will once they are unemployed). More importantly, it turns out you don't necessarily need Vertica for fast out-of-core loading and processing.

Granted, there are plenty of other ways to work out of core (hdf5, bigMatrix, any random database, blah blah) but this was one that was new to me. And I like it.

postgresql used to have a way to embed R into PostGreSQL too:

https://github.com/jconway/plr

I guess it still does, though I haven't used it in years. You can, of course, do the same thing with Python.