{"id":847,"date":"2010-03-01T12:05:00","date_gmt":"2010-03-01T12:05:00","guid":{"rendered":"http:\/\/scientopia.org\/blogs\/goodmath\/2010\/03\/01\/animal-experimentation-and-simulation\/"},"modified":"2010-03-01T12:05:00","modified_gmt":"2010-03-01T12:05:00","slug":"animal-experimentation-and-simulation","status":"publish","type":"post","link":"http:\/\/www.goodmath.org\/blog\/2010\/03\/01\/animal-experimentation-and-simulation\/","title":{"rendered":"Animal Experimentation and Simulation"},"content":{"rendered":"

In my post yesterday, I briefly mentioned the problem with simulations
\nas a replacement for animal testing. But I’ve gotten a couple of self-righteous
\nemails from people criticizing that: they’ve all argued that given the
\nquantity of computational resources available to us today, of course<\/em>
\nwe can do all of our research using simulations. I’ll quote a typical example
\nfrom the one person who actually posted a comment along these lines:<\/p>\n

\n

This doesn’t in any way reduce the importance of informing people about
\nthe alternatives to animal testing. It strikes me as odd that the author of
\nthe blogpost is a computer scientist, yet seems uninformed about the fact,
\nthat the most interesting alternatives to animal testing are coming from that
\nfield. Simulation of very complex systems is around the corner, especially
\nsince computing power is becoming cheaper all the time.<\/p>\n

That said, I also do think it’s OK to voice opposition to animal testing,
\nbecause there *are* alternatives. People who ignore the alternatives seem to
\nhave other issues going on, for example a sort of pleasure at the idea of
\npower over others – also nonhumans.<\/p>\n<\/blockquote>\n

I’ll briefly comment on the obnoxious self-righteousness of this idiot.
\nThey started off their comment with the suggestion that the people who are
\nharassing Dr. Ringach’s children aren’t really<\/em> animal rights
\nprotestors; they’re people paid by opponents of the AR movement in order to
\ndiscredit it. And then goes on to claim that anyone who doesn’t see the
\nobvious alternatives to animal testing really<\/em> do it because they
\nget their rocks off torturing poor defenseless animals.<\/p>\n

Dumbass.<\/p>\n

Anyway: my actual argument is below the fold.<\/p>\n

<\/p>\n

So – what’s the problem with simulation? It’s not that we don’t have
\nenough computational power. The amount of computational power available to us
\nis truly mind-boggling. Many things that would have been completely impossible
\njust a few years ago have become absolutely routine. Sitting on my lap, I’ve
\ngot a computer with two CPUs, each running at 2.5 gigahertz. Under my desk,
\nI’ve got a computer with 8 CPUs. And I routinely run programs that use
\nseveral thousand<\/em> CPUs in a datacenter. That’s not unusual
\ntoday: people who want to run simulations can easily and cheaply
\nget access to clusters of thousands of computers.<\/p>\n

So computer power isn’t a problem. Using modern computers
\nand software, we’ve easily got enough power to run incredibly complex
\nsimulations. <\/p>\n

And that doesn’t help with using simulations to replace animal testing
\nat all<\/em>. The problem with using simulations for testing has absolutely
\nnothing to do with computational power. No matter how much computational
\ncapability you have available, there’s still a huge, fundamental problem with
\nsimulations. It’s not particularly hard to understand: simulations only do
\nwhat you tell them to<\/em>.<\/p>\n

Let’s start at the beginning: just what is a simulation? It’s a
\nmodel<\/em> of a real system, which attempts to reproduce the effects
\nand\/or behavior of the system that it models. In the case of computer
\nsimulations, which is really what we’re talking about, we produce a
\nmathematical model and algorithmically describe how the model evolves over
\ntime. In other words, we write a computer program that runs a mathematical
\nmodel of a process.<\/p>\n

And right there is the problem. We<\/em> produce the model, and
\nwe<\/em> implement the model. It does exactly what we tell it to. We
\nprogrammers have a saying which applies particularly well to simulations:
\ngarbage in, garbage out. Computers do exactly what you tell them to; if what
\nyou tell them to do isn’t right, then no amount of computer power is going
\nto change the fact that you didn’t tell them to do the right thing.<\/p>\n

If we really understand what we’re simulating, we can do simulations that
\nare astonishingly accurate. For example, we can do wind tunnel simulations of
\naircraft designs, and the results of those simulated experiments are perfect
\nto within our ability to measure them. We’ve got it down well enough that we
\ncan get more precise<\/em> results with a computational model than we can
\nwith a scale model in a wind tunnel. We understand how air flow works. We
\nknow how to model it. It takes a whole lot of computation to do a decent job
\nof it, but as I said before, lack of computational resources isn’t generally
\na problem.<\/p>\n

But we can’t simulate something if we don’t know how it works. And that’s
\nthe problem with biological simulations: we don’t know<\/em> how the
\nbiological systems work. If we don’t know how they work, we can’t build
\nan accurate model. And if we don’t have an accurate model, we can’t build
\nan accurate simulation.<\/p>\n

When it comes to biology, we just don’t know enough to
\nbe able to accurately or even meaningfully simulate many simple
\nprocesses.<\/p>\n

Let me give you an example. In 1956, a scientist named Otto Warburg
\ndiscovered that most cancer cells have an abnormal metabolism. Normal cells
\nmetabolize sugars to produce energy using their mitochondria. Cancer cells
\ndon’t usually use their mitochondria – they use an entirely different
\nmetabolic pathway. It’s called the Warburg effect<\/em>. (My
\nfriend and blog-father Orac sent me a note to say that the Warburg effect was
\nactually discovered in 1906 – fifty years earlier than the citation I found would suggest. So instead of having 50 years to study it as I write below, it’s over a hundred!)<\/em><\/p>\n

We’ve known about the Warburg effect for over fifty years. And
\nyet, we still don’t know why<\/em> cancer cells do it. We don’t
\nknow what causes it. There’s no way that we can simulate it. We can’t
\nwrite a simulation of a cancer cell, because we don’t have a sufficient
\nunderstanding of its metabolism: we don’t know when it’s going to use
\nWarburg. And the metabolism is one of the simpler<\/em> aspects of
\ncancer behavior: things like how the DNA in the cell changes in
\ncancer is vastly<\/em> more complicated. We’re just no where
\nclose<\/em> to understanding it – which means that we’re
\nno where close to being able to simulate it.<\/p>\n

So we can’t simulate it. Or, to be more precise, we can’t
\nproduce a simulation that we know is valid. We can’t simulate a process
\nthat we don’t understand. We can’t simulate a process that we don’t
\nknow about. And biology and medicine are just chock full of processes
\nthat we don’t understand, or that we aren’t ever aware are occurring. So
\nwe can’t simulate those.<\/p>\n

There’s another aspect of simulation that’s important: validation. Even in
\nthe best simulations, you can’t be sure that the simulation is correct until
\nyou test it. That testing is called validation<\/em>. To validate a
\nsimulation, what you need to do is to take some starting point in both the
\nreal world and in the simulation, and observe both for some period of time,
\nand then check that the results of the simulation and the real-world match.
\nFor example, for air-flow simulations, you put an object into a wind-tunnel and
\nmeasure everything; then you simulate putting an object into a wind-tunnel
\nwith the program; and then you campare the results.<\/p>\n

Without validation, you have no way of knowing if your model is
\ncorrect.<\/p>\n

Validation is really key when you don’t understand what’s going on. I’ll
\npull out another example. I come from a family with a history of clinical
\ndepression. I suffer from it myself. For me, and many other people, there’s a
\nclass of drugs called selective serotonin reuptake inhibitors (SSRIs). SSRIs
\nwere developed based on a theory about what caused depression – specifically,
\nthat depression was caused by a shortage of the neurotransmitter serotonin in
\nthe brain. So they developed a set of highly targeted drugs that muck with the
\nserotonin chemistry of the brain. And lo and behold, it works<\/em>. It’s
\noften cited as an example of the amazing success of modern targeted
\npharmaceutical development.<\/p>\n

Except that all of the most recent research about depression appears to
\nshow that the serotonin theory of depression is wrong<\/em>. And yet, for a
\nsignificant number of people, the drugs work – and they work dramatically
\nbetter than you’d expect from placebo effects. They do<\/em> increase the
\navailability of serotonin in the brain. They also do a bunch of other things –
\nlike stimulate cell growth in certain parts of the brain. How do we figure out
\nhow SSRIs work? Among other methods, computer simulation with validation from
\nanimal models. We start with a model: what would we expect to see if the
\nserotonin model were correct? We work out that model in detail. We convert
\nit to a program that predicts what kinds of biological effects we should
\nexpect to see if it’s correct. Then we do the test on a group of animals,
\nand check to see if what we observe in the animals matches what we expect
\nfrom out model. Based on the results of that, we can judge how well the model
\nmatches reality.<\/p>\n

From all of this, it might sound like computer models are boring and not
\nterribly useful. After all, all that they can do is what we tell them to. So
\ndon’t we always know the results before we start?<\/p>\n

Fascinatingly, no. We have some amazingly precise models of physical
\nphenomena, but understanding what those models mean when they’re scaled up to
\nlife size is incredibly difficult. And in fact, sometimes, we simply don’t
\nknow how to see what a model says about how a system will evolve through time
\nwithout actually watching it progress. For example, I mentioned computational
\nfluid flow above. We don’t know if it’s possible<\/em> to compute the
\nresult of a fluid flow system without simulating it through discrete time
\nsteps. Simulation is incredibly useful, because it lets us take dynamical
\nsystems, and watch how they evolve over time. <\/p>\n

So simulation absolutely has a role. And hopefully, it can reduce<\/em>
\nthe number of animals that are used in medical experimentation. But it can
\nnever replace<\/em> them. There really is no substitute for reality.<\/p>\n","protected":false},"excerpt":{"rendered":"

In my post yesterday, I briefly mentioned the problem with simulations as a replacement for animal testing. But I’ve gotten a couple of self-righteous emails from people criticizing that: they’ve all argued that given the quantity of computational resources available to us today, of course we can do all of our research using simulations. I’ll […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"jetpack_post_was_ever_published":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":false,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2}},"categories":[7,79],"tags":[],"class_list":["post-847","post","type-post","status-publish","format-standard","hentry","category-bad-software","category-computation"],"jetpack_publicize_connections":[],"jetpack_featured_media_url":"","jetpack_shortlink":"https:\/\/wp.me\/p4lzZS-dF","jetpack_sharing_enabled":true,"jetpack_likes_enabled":true,"_links":{"self":[{"href":"http:\/\/www.goodmath.org\/blog\/wp-json\/wp\/v2\/posts\/847","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.goodmath.org\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.goodmath.org\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.goodmath.org\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.goodmath.org\/blog\/wp-json\/wp\/v2\/comments?post=847"}],"version-history":[{"count":0,"href":"http:\/\/www.goodmath.org\/blog\/wp-json\/wp\/v2\/posts\/847\/revisions"}],"wp:attachment":[{"href":"http:\/\/www.goodmath.org\/blog\/wp-json\/wp\/v2\/media?parent=847"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.goodmath.org\/blog\/wp-json\/wp\/v2\/categories?post=847"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.goodmath.org\/blog\/wp-json\/wp\/v2\/tags?post=847"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}