Mathematics and Biology

For the discussion of the sciences. Physics problems, chemistry equations, biology weirdness, it all goes here.

Moderators: gmalivuk, Moderators General, Prelates

User avatar
Velifer
Posts: 1132
Joined: Wed Dec 26, 2007 4:05 pm UTC
Location: 40ºN, 83ºW

Re: Mathematics and Biology

Postby Velifer » Mon Nov 16, 2009 2:07 pm UTC

CrazyIvan wrote:You are however locking yourself out more advanced "Biology related" fields, like Mathematical Biology, Epidemiology, etc. which are fairly math intensive. Often you'll have extremely Math-oriented collaborators, but it is useful enough to understand what the hell they're talking about.


Field epi is decidedly non-math. 2x2 tables and odds ratios-- the workhorse methods of the county health department epi-- are not the stuff of math journals. Those bar napkin calculations work. They are typically enough to handle a food borne illness outbreak. You can be a very good epi with very simple tools and a basic understanding of math.

Research is a bit different, but even on the social side where it is so tempting to make elaborate models with heroic assumptions, the best models are extremely simple. The danger is that people try to use tools in new ways without understanding how they work. That is where knowing bio and being strong on the math side makes a difference. It works both ways. Some of the worst errors I've seen were made by mathematicians who were consulting/collaborating but didn't know a damn thing about bio.
Time flies like an arrow, fruit flies have nothing to lose but their chains -Marx

CrazyIvan
Posts: 30
Joined: Thu May 28, 2009 4:17 pm UTC

Re: Mathematics and Biology

Postby CrazyIvan » Mon Nov 16, 2009 9:36 pm UTC

Velifer wrote:
CrazyIvan wrote:You are however locking yourself out more advanced "Biology related" fields, like Mathematical Biology, Epidemiology, etc. which are fairly math intensive. Often you'll have extremely Math-oriented collaborators, but it is useful enough to understand what the hell they're talking about.


Field epi is decidedly non-math. 2x2 tables and odds ratios-- the workhorse methods of the county health department epi-- are not the stuff of math journals. Those bar napkin calculations work. They are typically enough to handle a food borne illness outbreak. You can be a very good epi with very simple tools and a basic understanding of math.

Research is a bit different, but even on the social side where it is so tempting to make elaborate models with heroic assumptions, the best models are extremely simple. The danger is that people try to use tools in new ways without understanding how they work. That is where knowing bio and being strong on the math side makes a difference. It works both ways. Some of the worst errors I've seen were made by mathematicians who were consulting/collaborating but didn't know a damn thing about bio.


To quote myself: "So, having picked up some basic stats, you're set for some microbiology and ecology research, teaching, applied public health, etc."

"Epidemiology" as a field description, like Biology, is nigh useless. Field epidemiology, indeed, can be done largely with 2x2 tables, odds ratios and the like. Although I'd argue that the moment you start using chi-squared tests you may very well have violated hypothetical undergrads "No Math" rule.

The problem is from there it gets worse, from a math perspective. Epidemiology as a field, outside of applied/field Epi, is getting more methodologically complex, not less.

But then, I'm biased. I've gotten in several arguments with more mathematically inclined folks who labor under the belief that "Epidemiology" stops at a 2x2 table.

User avatar
Velifer
Posts: 1132
Joined: Wed Dec 26, 2007 4:05 pm UTC
Location: 40ºN, 83ºW

Re: Mathematics and Biology

Postby Velifer » Tue Nov 17, 2009 3:52 pm UTC

CrazyIvan wrote:But then, I'm biased. I've gotten in several arguments with more mathematically inclined folks who labor under the belief that "Epidemiology" stops at a 2x2 table.

I get into arguments with epis who think the same way. Learn the fucking methods, people! Biologists (in that very broad sense) do this to themselves. "Oh no, we don't need to know that... that's math. Leave it to the guy helping us with the stats!"
I've heard that a few times. It makes me feel all stabby.
Time flies like an arrow, fruit flies have nothing to lose but their chains -Marx

User avatar
Izawwlgood
WINNING
Posts: 18686
Joined: Mon Nov 19, 2007 3:55 pm UTC
Location: There may be lovelier lovelies...

Re: Mathematics and Biology

Postby Izawwlgood » Tue Nov 17, 2009 5:19 pm UTC

Velifer wrote:"Oh no, we don't need to know that... that's math. Leave it to the guy helping us with the stats!"
I've heard that a few times. It makes me feel all stabby.

Why? The point of science collaboration is to bring together people who don't have the time or the expertise to master everything required for their discovery. You should never limit yourself or your learning, but if tasked with writing a huge piece of code to do something, it'd be way more time efficient for me to simply ask a physics grad student for help, then to learn matlab, try and figure out how to write the code, and troubleshoot it. Which isn't to say I don't plan on learning matlab.
... with gigantic melancholies and gigantic mirth, to tread the jeweled thrones of the Earth under his sandalled feet.

CrazyIvan
Posts: 30
Joined: Thu May 28, 2009 4:17 pm UTC

Re: Mathematics and Biology

Postby CrazyIvan » Tue Nov 17, 2009 8:31 pm UTC

Izawwlgood wrote:
Velifer wrote:"Oh no, we don't need to know that... that's math. Leave it to the guy helping us with the stats!"
I've heard that a few times. It makes me feel all stabby.

Why? The point of science collaboration is to bring together people who don't have the time or the expertise to master everything required for their discovery. You should never limit yourself or your learning, but if tasked with writing a huge piece of code to do something, it'd be way more time efficient for me to simply ask a physics grad student for help, then to learn matlab, try and figure out how to write the code, and troubleshoot it. Which isn't to say I don't plan on learning matlab.


For the same reasons the math collaborator should be relatively up to speed on the biology in question. Your research should never, ever be a black box. Sure, the physics graduate student can help hammer out the code in matlab, but you should know what they're doing. Similarly, you should never hand over your data to "the statistics guy" and say "do stuff with it, lemme know how it turns out". Because that's not collaboration.

User avatar
Mokele
Posts: 775
Joined: Fri Aug 21, 2009 8:18 pm UTC
Location: Atlanta, GA

Re: Mathematics and Biology

Postby Mokele » Tue Nov 17, 2009 9:07 pm UTC

sure, the physics graduate student can help hammer out the code in matlab, but you should know what they're doing.


The problem with this topic is that it's not a binary between complete understanding and complete ignorance. Being completely uninformed is obviously bad, but absolute, complete understanding may take far more time and effort than is worth it.

I'll use an example from my own recent work. I wanted to measure some distances moved by animals in a test arena (without interrupting), but was restricted in terms of camera position and number. I hit upon the idea of doing a perspective transformation (since the animals would always be at roughly ground level when measured), but try as I might, I couldn't get my head around the math (and MATLAB's idea of 'help' is only helpful if you already have a degree in math). Eventually, I just used the MATLAB function to make a transform based on a calibration grid, and checked it by running the same procedure on randomly placed 7 foot sticks.

The key is that if I'd had to learn it all, I'd've failed, because this was fieldwork and I had highly limited time. But because I learned a *bit*, I knew when to use it (critter on the ground) vs when not to use it (critter jumping high into the air), and I checked using the sticks of known length.


In a sense, both CrazyIvan & Izawwlgood are right - you don't want to be totally in the dark, but you should also use collaborators to 'fill in the gaps' in your knowledge and abilities rather than spend all the time and effort to gain said abilities.
"With malleus aforethought, mammals got an earful of their ancestor's jaw" - J. Burns, Biograffiti

User avatar
iop
Posts: 930
Joined: Tue Jul 24, 2007 11:26 am UTC
Location: The ivory tower

Re: Mathematics and Biology

Postby iop » Wed Nov 18, 2009 1:19 pm UTC

Mokele wrote:I hit upon the idea of doing a perspective transformation

That's the key. You have sufficient understanding of geometry that you know what you're looking for, and you are able to figure out whether the result makes sense. That's how much, more or less, know about, say, structural NMR, and then they have someone put the sample into the machine and produce a structure prediction from the curves. The biologist knows when it may be useful to try to do structural NMR (though they may be surprised how large the molecules can be nowadays), and the biologists can tell whether a structure looks plausible. Furthermore, the fNMR people know enough about biology to be able to suggest to the biologist how to prepare the sample without demanding the impossible, and they can make suggestions as to the size of the protein fragment that may be needed to get the domain to fold.


What the complaining here is about that there are still too many occasions where the understanding is binary when it comes to bio and math - the biologist is not interested in math but has to get something "math-y" done, and the programmer (who, by the way, most likely uses libraries someone else has written - only fools write everything themselves) has no interest in or understanding of biology, and thus hacks up something that won't actually work with real data, and since the biologist doesn't know better, crappy code is used to produce crappy results and everyone's time is wasted.



By the way: I think Matlab's help is impressive. Sure, it doesn't teach you all the details of how to write a specific function yourself (though you can often actually look at the function if you really want to know), but it does teach you how it works. And the demos/tutorials are full of helpful examples.


Return to “Science”

Who is online

Users browsing this forum: No registered users and 13 guests