Almost everybody gets taught an over-simplified (but true!) version of the scientific method, which is that you
(2) invent one or more hypothetic mechanisms that could have caused the phenomenon; next, (3) design experiments that could prove or disprove whether each of your alternative hypotheses is true or false. And also (4) do these experiments.
Actual research might surprise you if you expect only these 1, 2, 3, 4 steps. Most working laboratories are trying to find more evidence or examples that confirm a hypothesis or combination of hypotheses that they strongly believe in, already. UNC gives undergraduate students many excellent examples to become part of active research laboratories, and for many students these experiences become the best and happiest parts of their education. Many or most such students actually make new discoveries. But when you do start research in a laboratory, don't be surprised if everybody in the lab strongly believes in certain theories (that they are so sure of, they regard them as common sense). In case you should invent some alternative theories, and have some ideas about how to disprove your lab-mates' favorite theories, please realize it will be heresy. Save your brilliance until you have working in that lab for a few weeks, and have succeeded in completing some of the experiments (or "experiments") that they assigned you to do.
A few heretical ideas are brilliant; and an even tinier fraction of these brilliant ideas may also turn out to be true! (99% of the smartest ideas may deserve to be true, but just aren't)
The most influential book on the Philosophy of Science in the last 50 or 100 years was by Thomas Kuhn ("The Structure of Scientific Revolutions", which is a very readable book). Kuhn himself did not believe that this meant that science "isn't really true"; but that was how his writings have been interpreted by tens of thousands of "Post-Modernists", "Deconstructionists", Sociologists and Anthropologists (& Many other 'ists' too numerous to mention). You probably will run into this view in some of your liberal arts courses. Don't be shocked by it. Most research scientists who have actually read Kuhn's books (me for example) agree with nearly all Kuhn's ideas, and we know that he didn't really say what the "Post-Modernists" etc. claim that Kuhn proved. (Really, they are post-Marxist nihilists) Don't let them get you down; let their claims stimulate your thinking without giving in to b.s. Kuhn said that even the best scientists cling to the favorite theories of their time, and rarely abandon them until
AND (b) Some scientist (often from another field) invents and popularizes a new theory, that DOES explain (=predict?) the data that the old theory couldn't predict. They won't give up the old theory without a better new one; and they tend to go down fighting and sneering at the new theory (until eventually they decide that they knew it all along).
The progress of science is usually a zig-zag. It moves toward the truth like a trail going up a steep mountain. That doesn't mean that the top of the mountain doesn't really exist, "because if it did, then trails would always point directly toward the top." Post-modernists are a little like "Pit-Preachers", and eventually you will encounter some of them, probably teaching some course, and in a position to grade your responses to their claims! Don't be too shocked when this happens: you can still learn a lot from people who are wildly wrong. One research laboratory might concentrate on evidence that protein phosphorylation controls cell growth, or that microtubules pull chromosomes, or that chemical diffusion gradients inside embryos control which organs form at any given location, etc. The other most influential philosopher of Science of the 20th century was named Karl Popper, and his key idea was that theories are not really scientific unless they make reasonably specific predictions about what should NOT be possible: in other words, what observations could be made that would disprove the theory. If a theory can "explain" any possible result or observation, then Popper regarded it as pseudo-science (not necessarily false; but not scientific!) Popper's main targets were Marx and Freud (and communism, and psychotherapy, etc.) And also "Creationism"! In fact, for several years Popper also argued that Darwinian Evolution was not scientific (!) on the grounds that no matter what genes or fossils anyone discovered, some evolutionist would surely invent a plausible story about how mutations and natural selection produced those results. Biologists eventually talked Popper out of this conclusion, by giving him lists of logically-possible results that they would regard as impossible to explain by evolution. This is a more interesting subject than you might expect; and Darwin's theory was confronted with several sets of logical and factual evidence that really COULD have disproved it (IF it hadn't been right), and that Darwin himself wrote about at length, criticizing his own theory. One example was that there didn't seem to be any way to explain the evolution of electric eels, because their ancestors must have gone through intermediate evolutionary stages that could only produce 1 or 2 volts (before getting to 100 to 600 volts!), and those low voltages wouldn't have conferred any advantage, in the sense that a fish that can generate 2 volts would not leave more descendants that a fish that could only generate 1 volt. Later, it was found that fish use weak electric fields to sense nearby objects (like radar, sort of), and sometimes as mating signals (like electric frog calls). So that explained the evolution of low voltages; and then when their "radar" got good enough, they could also use the voltage to shock enemies or kill food, or get jobs in the circus. (Just kidding on the last one.) Another objection was that if genes behaved liked mixing fluids, then favorable mutations would always get diluted out. This is a mathematically difficult argument; but is agreed by experts to be valid. So it turns out that natural selection wouldn't work well enough, except for the fact that genes have the discrete properties that are what Mendel discovered. Not one Biologist in 25 realizes this, as far as I can tell. If pro-Darwin Biologists had realized that Mendel's discoveries provide an escape from this mathematical "disproof", then those discoveries would have gotten the attention they deserve much sooner! In fact, Bateson and some of the other re-discoverers of Mendel reasoned that his results somehow disproved Darwin's theory. Bateson was just wrong about this.
"Publication"The one most important scientific method (in the last 150 years), is publication of results and conclusions in "refereed journals". Journals are those thick, magazine-looking things in science and math libraries. When a scientist (or a research mathematician) believes that he or she has made a really good discovery, what they do is to write a complete description of what they have seen and done regarding this discovery, probably with photographs and graphs and diagrams (and these days sometimes with a video disk containing movies of the phenomenon), and also including a list (bibliography) of relevant papers that have been published in the past, whose contents either support of contradict your conclusions and methods. Such a manuscript goes into fanatical detail.Next you give rough copies to friends (and also enemies) for them to criticize and find all the mistakes and weak points. Then you mail (or e-mail), or Fed-Ex, 3 or 4 duplicate copies of your manuscript to the editorial offices of the most highly-respected scientific journal that you think might possibly be willing to publish what you have written. "Nature" and "Science" are the two most prestigious journals, and have been for about the last 60 years. But there are thousands of different journals. This process is a little like applying to college; and not everyone who gets 1500-1600 on their SATs applies to Harvard; and not every terrific discovery gets submitted to Nature. But there is a temptation to see if you can get in. When Science rejects your manuscript, then you can re-submit it to the Journal of Cell Biology, or somewhere. One contrast to college applications is that it is not ethical to submit the same research discovery to more than one journal at a time. When you submit, you sign a letter promising that the manuscript is not simultaneously been submitted to any other journal. So people often prepare another version of the manuscript, to send off to Science ten minutes after they get the rejection letter from Nature. "Refereeing" is how the journals decide which papers to publish and which ones to reject. Science and Nature both reject well over 90% of the papers that get sent to them, and other good journals also reject 90%, 80%, or something like that (even though any sensible scientist will only send them there very best papers.) Ambitious scientists submit each paper to the most competitive journal where they think that paper has any reasonable chance of being accepted. Then when it's rejected, you send it to a less competitive journal, etc. Eventually, you can get it in somewhere; although I once gave up on a manuscript I thought was very good, but nobody else liked. For each research manuscript that is submitted to them, the editors of "refereed journals" then choose one or two or three other scientists who they think are the best experts on the general subject (almost always they choose two). The editor then sends copies of the manuscript to each of these experts, and asks them "to referee" or "to review" the manuscript. That means for them to write the most careful criticism of the manuscript that they can, including its good points and its bad points, and to advise the journal editor if the manuscript is good enough to be published (and whether readers of that journal would probably be interested in the subject). These reviewers are anonymous, in the sense that the authors of the manuscript are not told what other scientists wrote these criticisms, and recommended acceptance or rejection. To do a good job referring a manuscript can easily take one or two days (all day, as hard as you can think!). A scientist puts his or her reputation on-the-line every time they review a manuscript. After all, the editor knows who you are; and even if the authors of the manuscript don't know, they can usually make a good guess. Trying to figure out who rejected your manuscript is a major indoor contact-sport, with many bruises and hard feelings. Often, I deliberately split infinitives in referee's reports, to throw off suspicion (because of a very excellent, no-nonsense high school English teacher who flunked any paper that contained even one split infinitive. He said, "they aren't "wrong", but you should realize when you split one".) I thereby learned, never to split an infinitive. (= to never split, etc.) If one referee says the paper is good, and the other says it's bad, then they send it to some third expert (and sometimes the first two referees are asked to agree between themselves on who the third referee should be, who will break the tie). That is very flattering, to be chosen as the third person by the first two; but then you have to spend a solid week, sitting up late, writing the most incredibly careful report you can. In a small way, it's like being a judge, or something. If you read "acknowledgements" sections at the ends of published papers, sometimes you notice statements like this "The authors also wish to thank one of the anonymous reviewers for suggesting the fourth experiment reported in this paper. That experiment (etc.) had not previously occurred to any of the authors." If both referees are sufficiently favorable, then the journal publishes the manuscript. Often changes are required, such as when they tell you to cut the text to 60%, take out at least 3 photographs, and entirely leave out one or two of your conclusions, because the reviewers and the editor don't think your evidence really proves the truth of everything you claimed. Most journals mail printed copies of both reviewers' comments not only to the authors of the paper, but also to the reviewers themselves. That way you can remember what you wrote, and compare it with the criticisms of the other reviewer. After a paper is accepted it can take 6 months or more for it finally to get published. During this time, the authors are like blushing brides, flitting to the library 3 times a day to look for the latest issue of that journal. Finally, it appears in print, and you get congratulated by others in the department. You may have heard the phrase "Publish or Perish", and for research scientists, what matters is publication in refereed journals. The more competitive the journal, the more the brownie points. Some brownie points also come from having been a referee for the best journals, and some of them publish a list at the end of each year: "The following scientists have served as anonymous referees for this journal between Jan of 2004 and Jan of 2005" or something like that. Serving as an editor of a journal earns you even more credit. To be completely blunt, publication in refereed journals is the key criterion by which people get appointed to be Professors, and get promoted from Assistant Professor to Associate Professor to Full Professor, and to special higher titles like "Kenan Professor". Teaching also matters at UNC, in the sense that professors won't get promoted if their lectures aren't clear enough, and accurate enough. But unless and until we publish several major discoveries in at least some of the most competitive journals, we have no chance at promotion or tenure. Coaches whose teams don't win enough games get the same treatment as Assistant Professors who don't publish enough refereed papers. Other sorts of criteria are used in other professions. Sometimes there is a very difficult test, like the Bar Exam, that you have to pass to become a lawyer. Historically, the Ph.D. degree was invented as being awarded in return for some significant new discovery, which you are supposed to 'Publish'; and beginning in Germany in the mid 1800s, and spreading to England and then to the US in the late 1800s, publication of new discoveries became the criterion for hiring and promotion of teachers at the level of colleges and universities. This was a big change from earlier methods, and in each country that adopted this method, it was immediately followed by revolutionary improvement in scientific productivity of universities. It sounds like an exaggeration to suggest that the modern world is largely the result of using refereed publication as the criterion being a professor (more than any one other cultural change). Anyway, this was one of the major historical changes of modern times. Also, any time you hear some claim that sharks don't get cancer, or that shark soup cures cancer, then the question to ask is: "Was that published in a referred journal?" If it wasn't, then it doesn't make up for it that maybe Mike Wallace said it on "60 Minutes". Anybody that has half-way good evidence for any important theory will submit that evidence to refereed journals of some quality. Mistakes sometimes get published, especially in not-too-competitive journals. And even Science got hoodwinked a couple years ago. But if a scientist doesn't publish what he discovers, he can't be too serious about it; and there is no reason to believe it, whether or not it is on television. Sometimes very revolutionary discoveries get unfair reviews, but publication is the best way anyone has of proving the truth.
Model organisms:Most research is becoming concentrated on a relatively small list of carefully chosen plants, animals, bacteria and viruses. The fruit fly "Drosophila" has been one of these "model organisms" for almost a century, and so has the house mouse. A small plant named Arabidopsis and a round-worm named C. elegans have become two of the most intensively studied organisms on earth.The research strategy is to concentrate research on a few species, until "everything" possible is learned about each of these particular species; and after that, general conclusions can be reached that will (certainly?) be true of all related species. In this department, the majority of the professors' laboratories concentrate on some particular model organism. There are fly labs, yeast labs, C. elegans worm labs, Xenopus frog labs, and a lot of Arabidopsis labs. A couple of labs concentrate on particular methods (optics; computers), and some study a broad category of phenomena. But the latter are becoming more and more the exception. Good properties for a "model organism":
Short life cycle, in the sense that it only takes a month for eggs to hatch and grow to the age where they can lay more eggs. If one species takes twice as long to produce a new generation, then you can only do half as many experiments in a given amount of time. Small genome, relative to other species in the same taxonomic category (for example, other frogs; or other algae) Not cute; so you won't mind killing millions of them: because you will probably have to.
Genetic screens(method for finding mutant organisms in which the altered genes code for proteins that function in some normal process in which you are especially interested.)Suppose you want to discover all the proteins that participate in, for example, pumping ions into cells: how could you find them. Here is a general method: Raise millions of C. elegans worms; then expose them all to radiation, or to chemicals that are known to damage DNA. Following that, you need some way to find the "needles in the haystack" worms that have been mutated in any of the genes that code for proteins that pump ions. Those worms are going to be at least somewhat abnormal in their ion pumping ability. So maybe they will tend to be paralysed in high concentrations of some salt (more than non-mutant worms, and also more than worms that have mutations that code for proteins that serve other functions). Or maybe they will wiggle more in response to salt; or maybe they will survive longer in medium with very low salt concentrations. Any of these are possible, and researchers have to use their imagination trying any conceivable way to find just those few worms that have mutations in the kinds of gene you are interested in. The ideal "screen" would be some treatment that killed all the worms that do NOT have a mutation in at least one of the genes for proteins of the kind you want to discover. If you can kill all the others, or make them crawl away leaving behind only the kinds of mutant you want to study, then even if only one worm in fifty billion has such a mutation, then you can find those tiny few. Having found them, then you can grow up millions of children of those mutant worms, and at last get down to studying whichever proteins in them serve in ion pumping, or whatever you are interested in. You already know the concept of the "resolving power" of microscopes. So can you also see that some genetic screens are better than others, where the meaning of "better" is that they can help you find more needles in bigger haystacks, and (just as important) NOT find too many objects that are not really needles, but look like needles. Logically, "genetic screens" are equivalent to sifting something out, and sometimes you really do sift, with some kind of literal screen or filter. Other times, the separation is by the death or survival of non-mutants versus mutants, or by their movement, or change in color followed by picking the mutants out with tweezers. It is always the mutants that you want to survive, or somehow to separate themselves. It's not such a good screening method if the mutants die and the non-mutants survive. You need living mutants; and you need just those tiny fraction of the mutants in which the altered gene codes for a protein that pumps ions, or whatever process you want to study. Scientists sometimes invent the most fiendish screening methods for finding particular kinds of mutants. More than one Nobel Prize has depended directly on a devilishly clever genetic screen. When they announce the prize, they say "for discoveries of fundamental mechanisms by which proteins control cell survival" or something like that. They don't say "This prize is in honor of inventing a method for isolating cross-eyed flies, with minimal contamination from flies that are merely near-sighted." But that may have been the key discovery. If you volunteer to do research in some biology dept lab, it is likely that part of what you do will be finding individual worms or yeasts or something, that appear to have mutations that affect some particular function. Another example of a genetic screen (presented in lecture) The green alga Chlamydomonas is a single cell that swims by using two flagella. The structure of these flagella is almost exactly the same as that of mammalian sperm cells and cilia, so this is a very useful model organism for studying male infertility and diseases resulting from defective cilia, for example polycystic kidney disease. To learn more about the proteins that are involved in flagellar structure and function, you can do a genetic screen to find mutants that can't swim, and then study these using genetics, microscopy, biochemistry and molecular biology to determine what genes (and hence proteins) have been altered.
2) Grow the population of cells in a tall container with a light at the top. The normal cells will swim toward the light, and the ones that can't swim will collect at the bottom of the container. 3) Collect the cells that can't swim, dilute the culture, and spread them on agar, so that each starting cell in the population will give rise to a separate colony of cells 4) Analyze each of these colonies to determine what seems to be wrong with their swimming. For example, some might make flagella that are too short, or no flagella at all. Others might make flagella that are normal length but can't bend. Still others might have flagella that are uncoordinated, so that although the flagella can move, the cell can't swim forward. This much can be done at the light microscope level. 5) Then for each population, characterize the defect further: perhaps a structure looks abnormal when examined by electron microscopy, or one or more proteins are missing when all the proteins of the flagella are separated by electrophoresis. You can also cross one mutant to another to see if they can have normal offspring (which would suggest that the mutations are in different genes). 6) You might then try to transform the mutant cells by bombarding them with DNA from normal cells to see if you can "cure" the defect. By subfractionating the normal DNA, you can gradually zero in on a short piece of DNA that contains the normal version of the gene that has been mutated, and then sequence this DNA to deduce the exact amino acid sequence of the protein it encodes.
Study Questions 1) Briefly state the version of "The Scientific Method" that our textbook describes. 2) Contrast the textbook's scientific method with what these web pages say. 3) What did the philosopher Thomas Kuhn write about what most scientists actually do most of the time, instead of the classical "scientific method"? *4) What do many avant garde post-modernists believe that Kuhn proved? (usually they write "demonstrated" instead of "proved", I have noticed)
5) Kuhn concluded that most scientists will give up their favorite old theory and switch to a new theory only when what two things have both happened? **6) What is the best way to deal with a situation where you are in some course where the person doing the grading claims that Kuhn demonstrated that science isn't really true, but is just ideas that scientists have guessed, and cling to fanatically.
b) tell them "that's not what my Bio 11 teacher told us" (very bad strategy) c) ask whether Kuhn himself ever concluded that science is not the closest approximation to the truth that people have achieved so far (another bad strategy - unless maybe you save this question until after final grades have been posted)
When in Istanbul, don't make jokes about whether their computers run Unix. 7) When a new theory is published, do scientists in the field check to see whether it explains known facts as well or better than the theory that is currently accepted as true (according to Kuhn)? 8) Karl Popper's big contribution to the philosophy of science is what key idea, or claim?
9) According to Popper, theories are really "scientific" only if 10) What did Popper conclude about Marxism, Freudianism, creationism and even (for a while) Darwinian evolution? 11) What does it mean to say that some given scientific discoveries have been "published in a refereed journal"? 12) Who are the "referees" for scientific journals, and what job do they do? 13) Why do college professors try to publish their discoveries in the most competitive refereed journals that will publish them?
14) Is it considered ethical to submit the same discoveries to more than one journal at a time? *15) In many of the few cases in which someone has submitted the same discoveries simultaneously to more than one of the best journals, this has been detected by one or both of the referees, instead of by the editor. Explain why this makes sense.
16) Have college professors always been selected on the basis of research discoveries (since universities were founded in the late middle ages)? 17) What are some specific examples of "model organisms"? 18) Species are chosen to be model organisms on the basis of what characteristics? (name several)
19) The basic advantages of concentrating research on model organism species are what? 20) If you wanted to identify all of the proteins that contribute to some particular biological function, then explain how to use a "genetic screen" to find them. 21) If mutations in a certain gene cause abnormalities in a particular biological process, what can you conclude? *22) If even the smallest changes in a given protein cause the death of the organism, why does that make these proteins "invisible" to the genetic screen approach? *23) Does this depend on whether mutations in the proteins for these genes are dominant or recessive?
*24) Figure out approximately how to do a genetic screen specifically for recessive mutations that cause embryos to die at certain stages of development. *25) If mutating a certain gene consistently results in the death of embryos that are homozygous for this mutation, then what does that tell you about the normal function of the protein coded for by that gene?
|