The future was never now

/

Bioinformatics unavoidably requires prediction. There's only so much an analysis in silico will tell us about phenomena in vivo.

I'm worried that, as the availability of biochemical and genetic data becomes taken for granted, prediction will be replaced by meaningless, slavish futurism. It's already happening: a recent Wired magazine cover about a piece on CRISPR uses the headlines "No hunger. No pollution. No disease. And the end of life as we know it." Yes, it's hyperbolic. Yes, CRISPR-based technologies are already proving their use. It still falls into the same trap as so much other Wired-caliber futurism: the technology is advancing so rapidly that the futurism is already outdated. Unintended consequences can be addressed because they aren't some kind of unstoppable runaway train, they're a present topic of study.

In the Future, war is waged with levers. Also bombs, but that's nothing new.

In the Future, war is waged with levers. Also bombs, but that's nothing new.

A piece on Medium from a few months back labeled this kind of streamlined thought as futch (rhymes with brooch, not butch or crutch*).  The conventional wisdom has been that it's how we make big ideas accessible while retaining their freshly-polished chrome glint. Leave the gritty details to the scientists and engineers. If you're in one of those two groups, go interdisciplinary. Find someone outside your field to fill in those details you can't handle. Can't build the future yourself? Watch a TED talk! Join a Facebook group! Wait for the Future to land in your backyard!

The problem arises when the glossed-over details are those most relevant to everyday life. This is especially true of bioinformatics, as we've seen with issues like 23andMe's handling of their customers' sequence data.** Sure, personal genome sequencing seems like a hallmark of the Future, but it's here now and questions like "who owns my genetic data" are relevant to everyone with a genome.

*Or however Kanye pronounces it. It's just the vacay-form. Don't overthink it.

**See also: ethical issues concerning uBiome. Transparency is a good first step on a long path.

A few quick tips for better presentations and/or cutlery usage

/

Here's a tip: avoid presentations at all cost. Let someone else handle them.

Wait, no, hold on. That may be a quick tip but it's terrible advice. Everyone has to give presentations of some sort at some point in their career and/or personal life. They may be rigorously formal or may essentially be an improvised monologue at a party. When you have an idea to convey and an audience of more than a couple people,* you're giving a presentation.

Tip zero: Ensure lighting provides maximum silhouette effect. 

Tip zero: Ensure lighting provides maximum silhouette effect. 

In the sciences, our presentations usually take the form of Powerpoint and its ilk. I have to believe this software choice is mostly due to tradition rather than conscious choice. Using Powerpoint for every audiovidual-assisted presentation is like using a knife as your sole eating utensil: it'll get the job done but it'll never appear elegant.**

No matter the format, here are a few quick things to check in any presentation. They're all relatively minor and certainly not exhaustive. They're still easy mistakes to make, especially when you have to assemble a presentation with little time to spare.

  • Make sure to spell and pronounce every name correctly. You may get asked questions about the people attached to the names - or even by said people.
  • Thank those who've helped you but leave your thanks until the end.
  • Provide background early on in your presentation but never just a list of subjects. Context is key!
  • Briefly define any terms the majority of your audience isn't familiar with. Failing to do so can defeat the entire purpose of presenting.
  • If you've worked with some kind of a data set or database (and increasingly, you have, even if you're not giving a scientific presentation) be explicit about what the database contains and how it's useful. Even a frequently-used bioinformatics database like GenBank is often just referred to with phrases like "we got the sequences from GenBank." That's much like citing "the Internet" as a source in a research paper.
  • Similarly, don't talk about statistics without knowing what the statistics mean. That is, go beyond saying phrases like "the average of these values is 3.14" when you can tell the audience the context of that value. Is 3.14 higher than we'd expect? Even if it's statistically significant the context of other values, is it relevant? Did you just produce the value because it's what everyone else does and it's easy to do in Excel?

Here's the take-home message: Start programming a good replacement for Powerpoint now and your name will be lauded for centuries.

*You can even include animals in that audience if needed, though plants are only allowed if you're a magician.

** Some folks swear by Prezi but I've always found its zoom effects disorienting. Does that make me old?

Aged, hard cheese

/
“Edward Schreiber and Orson Anderson once tested whether the Moon really could be made of green cheese...Schreiber and Anderson compared the speeds of sound waves in rocks that were returned from the Moon with measured sound speeds of various terrestrial materials, including various types of cheese. (Sound speeds correlate highly with density and are thus often used to try to infer the composition of rocks.)”

The original paper (a copy, or a PubMed link if you really want it) includes measurements for a variety of Earth-bound minerals, moon rock samples, and seven cheeses, including the Swiss sapsago (a hard cow's milk cheese which is, due to added herbs, a bit green).

There's a lesson here: don't expect the newest hypothesis to explain everything, but don't blindly trust the old hypothesis, either.