Daniel Dvorkin @medigoth@qoto.org

Vive la Résistance!

A small entry in the "what is #bioinformatics, anyway?" category.

https://www.quora.com/What-is-the-difference-between-bioinformatics-and-other-computer-sciences-such-as-software-engineering/answer/Daniel-Dvorkin-3

This is one of those stories that either is the start of something huge or will sink like a stone: not much in between.

#Tessera's claims are ambitious but not unprecedented. One of our clients does something similar with stem cells, and they have a product that's actually made it to market. As far as I can tell, Tessera is claiming the ability to do this in all kinds of cells, with whole living organisms. Which would be a really big deal.

From a bioinformatics / biostatistics perspective (you knew that was coming) I have Questions about consistency. Doing it once, or ten times, or a thousand, isn't enough. You need to do it *trillions*, very accurately indeed. With cell cultures you just screen out the cells that didn't do what you wanted and keep the ones that did: insert "bury their mistakes" joke here. And with stem cells it's particularly easy to know when you have it right.

"65 percent efficiency—20 times the rate at which it performed unintended or errant edits" sounds quite good on a cell-by-cell basis, although I'd like to see the error bars on those numbers. In a patient, it would be an unmitigated disaster. Congratulations, we've cured your muscular dystrophy ... and given you ten different types of cancer.

More generally, their secrecy does not fill me with confidence, of the 95% variety or any other. As I've said many times before, science done in secret isn't science at all, it's alchemy.

Well. It's early days yet. Let's just hope Tessera isn't another Theranos, because I am *really tired* of that. If they do achieve any of what they're claiming, I'll be cheering them on.

#genetics #geneediting #genewriting #stemcell #bioinformatics #musculardystrophy #cancer

https://www.the-scientist.com/bio-business/can-gene-writing-deliver-what-gene-editing-can-t-70740

Seen in the wild: "All a #PhD tells me is you were at least able to cram study and pass tests."

I can't even.

Do people really think this? Obviously this guy does, but in general? An honest question, not a rhetorical one.

I've spent a significant chunk of my life in the company of people with #doctorates—including two grandparents, so I understood from childhood that getting a PhD was quite different from any other kind of degree. Most people I know haven't, but everyone who was part of my life through my long and winding #academic career has a pretty good idea what I did to become #Doctor Dvorkin. For anyone who doesn't know, I'm always happy to give a quick sketch.

Among people who don't have family or friends who have done it, and have never thought seriously about doing it themselves ... is that the perception, that grad school is just a more intense version of elementary school through college? More tests, more cramming, more regurgitation, and then you get some shiny new letters after your name?

If that's the case, I guess I can't blame them, exactly. Like I've said before, most people have no idea how most of the things I've done for a living in my life actually work. The #military, #medicine, #science: all quite mysterious except for what people *think* they know from movies and TV, which is often worse than complete ignorance. So I guess it's shouldn't be a surprise if #academia is on that list.

But I would really like to know how widespread this view is. Maybe I can do something with that.

Gondor has no knob. Gondor needs no knob!

[It's Time to Redeem Prometheus](https://www.esquire.com/entertainment/movies/a37377867/prometheus-movie-defense-essay/)

Not posted for agreement, if you share from my post please leave my commentary intact ... you know the drill.

The thrust of the article seems to be that the foolishness and total lack of relatability on the part of pretty much all characters makes sense in the light of Scott's message and vision for the movie. Making me not care about anyone played by Rapace and Fassbender, IMO two of the finest actors in the field, is an impressive trick, but he pulled it off!

In short: I recognize Scott had his vision. But given that it's a stupid-ass vision, I've elected to ignore it.

#Prometheus #Alien #Aliens #Xenomorph #RidleyScott #NoomiRapace #MichaelFassbender

"Free speech absolutist" must mean something different in Muskrat than in English.

The Scientist is a well-known and respected biomedical research trade publication, and I guess they put on events now and then. The Scientistt must be even better: that extra 't' adds a certain zing, doncha think? So I'd better send them my response right away. I'd hate to miss out on this opportunity.

Nice of them to recognize me as "a leading researcher in the field." Most of my work is behind the scenes, partly by personal preference and partly because it just doesn't take much to get the villagers lighting the torches. But maybe it's time for me to put aside my instinctive humility and step into the limelight.

Look. I understand why people are skeptical about the latest #fusion news from #Livermore. But it's too easy for skepticism to slide into cynicism, "fusion power has been twenty years away for sixty years" and all that. Don't gloss over the very real, steady progress researchers have made.

We know it's possible, and as long as we keep trying, someone will be the first to make it work.

Ever since the Human #Genome Project got rolling about thirty years ago (!) there's been a lot of hope, and a lot of hype, about "#personalized #medicine" or "#precision medicine." When it became clear that as always, the results weren't going to match the hype, a lot of the hope went away too. This is a mistake.

I'd like to talk about a quiet revolution in precision medicine: #genetic #dosage guidelines, a.k.a. #pharmacogenomic #labeling. The basic idea is that if you carry certain genetic #variants, you may need considerably more or less of a particular medication than the standard dose. Back in the '90s, the kind of genetic #analysis needed to make use of that information was far too expensive and time-consuming for #clinical practice. These days you can get a complete #sequence in a matter of hours, for the same cost as a battery of standard blood tests.

Fifteen years ago or so, the FDA approved the first pharmacogenomic labeling, for #warfarin. I was lucky enough to be in the room when the researchers made the announcement, and you could have heard a pin drop. Now it's routine, and there's a very long list: https://www.fda.gov/drugs/science-and-research-drugs/table-pharmacogenomic-biomarkers-drug-labeling

Everyone reacts to #medications differently. For most patients, most medications, and most diseases, there's a fairly broad range of clinical effectiveness between "too little to do any good" and "way too much." But for a substantial number of all of the above, the range is much narrower—and when you add up all the special cases, you get a hell of a lot of people!

A lot of #drugs never get approved, despite showing promise in clinical #trials, because they only help a portion of the study population. Regulatory bodies like the #FDA are notoriously resistant to #subgroup analysis, and I get why: it's very easy to cherry-pick those subjects in a clinical trial who happen to do well, and then come up with a post hoc explanation for why the test treatment worked for them but not for other participants. Some bad drugs have made it to market because of this kind of chicanery. But of course sometimes there's a real reason one group does better, and as long as genetic testing is part of the study design from the start, it's becoming possible to convince regulators that reason is valid.

My work is mostly upstream of this, in the drug #target #discovery phase: finding disease-related #genes and #proteins that might be modifiable with the right medication. Since it's part of the project from the start, that makes trial design easier, and the results more likely to be accepted. But I'd really like to see more #genomic analysis on drugs that *aren't* designed that way too, and I think we're getting there.

Genetic dosage guidelines, though, are making a real difference in current practice. There are still considerable debates over the merits of many labelings, driven partly by legitimate #statistical concerns and partly by ideology. But the principle is proven beyond reasonable doubt, and it's saving lives and relieving suffering right now, every day. Much more to come.

Note to self: when building a health survey which is going to be compared to an existing survey*, design your questionnaire to be as similar as possible to the existing one, even if it's not the way you'd do it. This will save you a lot of trouble down the road lining up all the data.

Oh well. Live and learn.

* #NHANES, in case you were wondering, which of course you were. https://wwwn.cdc.gov/nchs/nhanes/

I'm too tired to come up with any witty commentary about this. Just sad and disgusted.

Brief summary: paleontologist Robert #DePalma, who shares his cousin Brian's cinematic instincts, made a big spash a few years ago with the discovery of the "Tanis" site, which appears to preserve the day of the #Chicxulub impact. Since then he's milked it in the media for all it's worth. Well, as much as most scientists dislike that kind of behavior, it's how you keep the grant money rolling in. #Dinosaurs are maybe more popular than ever, but they're lousy at paying their own bills.

He's made many dubious claims before about the site, many carefully hedged so he could retract them later. It's a dangerous game. This time, it very much looks like he's done something much worse.

#Tanis is real, and it's a remarkable discovery. DePalma's inability to distinguish between flashy PR and outright fraud will accomplish nothing but tarnishing all the real #paleontology work done at the site, by anyone, for years to come. Good scientists like #During will suffer the consequences.

Yeah. Don't do that.

https://www.science.org/content/article/paleontologist-accused-faking-data-dino-killing-asteroid-paper

Too often things get worse. Sometimes they get better.

I continue to be amazed by the political shift in Muscogee County, Georgia, where with 67% of the vote counted, Warnock currently leads Walker 70%-30%. Of course it's possible that as votes come in from outlying areas, this lead will narrow. But Warnock seems almost certain to win the county handily.

Muscogee County is mostly two things, you see: the city of Columbus, named after you-know-who, and Fort Benning, named after a Confederate general of no particular note. (If they wanted a Georgian who *was* of note, they could have chosen Longstreet. Wonder why they didn't ... no, wait, I don't wonder at all.) When I was there, thirty-five years ago, it was still the CSA.

One of our drill sergeants was an Alabamian who grew up just across the state line. He warned us, "Now remember, boys, when y'all go out on pass ... around here you ain't in the Army. You in the *Yankee* Army." This was not a joke, and some of my barracks mates still have the scars to prove it.

Some years later, when I'd gone from green to blue, I had a friend at Minot AFB who had been born and raised in Columbus. He was a just enough younger than me, six or seven years, for there to be a meaningful generational divide. I told him that story, and he replied that although many people up to and including his older brother's age still thought that way, people his age and younger had rejected it. A bad old world, of which they wanted no part.

I was politely skeptical. Maybe, just maybe, I shouldn't have been.

Things can get better. Far too slowly. But they can.

Some answers just write themselves.

https://www.quora.com/If-evolution-is-legitimate-then-why-wont-evolutionists-adopt-baby-chimpanzees-and-teach-them-to-recite-Shakespeares-works-in-order-to-speed-up-the-evolution-process/answer/Daniel-Dvorkin-3

#creationism #evolution