BIO/CSC295 2011F, Class 02: What is Bioinformatics? Overview: * Dinosaur DNA, revisited. * The joys of interdisciplinarity. * Web exploration. Admin: * If you have not done so already, please fill out the RISC Survey asap. * We will start the Web exploration in class today. You should continue it on your own. You and your partner(s) should turn in answers to the Web exploration questions on Thursday. * No additional readings for Thursday! * EC for Thursday's "CS Extra" on HCI for Peace. 4:30 in 3821 Snacks in CS commons at 4:15 * EC for "Graduate School in Biology" Friday at noon In 2021, the large lecture room Bring a lunch, the College is a cheapskate * EC for Men's Soccer vs. Cornell at soccer fields Thursday at 5pm * One of only six home games this year (ten away) * Response papers are due * Email copies to both of us Why is this an "extraordinary paper"? * this = Asara et al. * Something never done before: No one had gotten protein sequences from fossils before * What is a fossil, anyway? * A bone * Well, what's left of the bone after organic material is eaten away * Hmmm ... proteins are organic molecules * Prof. P. says "Biological material turned to stone" * How can it be a fossil and still have peptides? * Boy, that's a problem * How do you preserve organic material for 68M years? * Well preserved * Lots of luck! * Great accident: Dissolve matrix (formerly organic) and magically there's some organic stuff left. [Remember: If Sam mistypes what you said, you should tell him.] * In reading this, you should have asked yourself "How, mechanicallistically, were the proteins preserved?" * The second paper talks about ways they avoided contamination * Summary * Got proteins from T. Rex bone through demineralization * Also ostrich (not fossilized) * Also mastadon (newer fossil) * Ran through mass spec to get sequences * Compare to existing sequences * You learn a lot from figures. What does figure 1 say? * It's a proof of concept. We're dealing with Ostriches rather than dinosaurs or mastadons to show that the technique works. * Part of their study is looking at possible changes over time. * Doesn't a mass spec just tell you what the sequence is? Why are we doing a proof of concept? * How a mass spec works * We use magnetic fields and ionization to find the mass/charge ratio of things * Break apart stuff into cations * The magnetic fields separate the cations into different mass/charge ratios * We can then determine how much we have of each ratio; this gives us a spectrum * So, for a particular molecule/peptide, you get a mass-spec spectrum * And the same peptide will always give you the same mass-spec spectrum * That's "same" in Bio terms, not "same" in CS. It's close, but not exactly the same * When we've built the spectrum for the T. Rex or Ostrich protein, what spectra are we comparing it to? * The Crichton technique: Start with mosquito in amber, sequence the DNA, find the genes, find the proteins, sequence them, do mass spec to build the spectra. * Whoops! Okay, there's no dinosaur DNA and there's no database * Question: What's a junk spectrum? * Prof P's data from the paper: First hit: Homosapien keratin type 1 Second hit: Homosapien keratin type 2. Hit number 4: BSA. Hit number N: e. coli * Conclusion: T. Rexs are closely related to lab techs. Except they are more friendly * Conclusion: The data are contaminated * But these contaminants may be normal when you're doing MS * And they are irrelevant to the dinosaur DNA * And maybe they ate chicken for lunch * But the match was not an exact match * And the chicken they serve at KFC and Chick Fil-A is mutated Review of where we are (and yes, we'll go back to the steps): * Reviewing the paper * Thinking about figure 1 * Stepping back to figure out their process (proof of concept) * Worrying about contamination Contamination * People who do mass spec are used to seeing this kinds of contamination * People who don't may criticize it * So that's why we don't like to release data * Controls (not discussed enough in the paper) * Another sample that does not contain fossil but that was handled identically otherwise * If there's chicken in the fossil, but not the soil sample, you have more confidence * Plain solution * Multiple treatments/preparations * Repeat the same experiment - multiple runs on multiple days * Repeat the same expermient at different labs * Not everyone should have a MS contaminated by ostriches * Replication with another sample from another dinosaur (or from multiple bones in the same dinosaur) * For some reason, curators don't like you to use up their fossils Mass spec: Compares spectrum from what you've gotten to a library of spectra. * Where's the database for dinosaur (or ostrich) (or mastadon)? * Try something that's close to dinosaurs (or ostriches) * Chicken * Newt * Frog * Crocodile - Not available * Alligator - Not available * Komodo Dragon - Not available * Hmmm ... isn't this starting with the answer? Why is the ostrich a proof of concept? * Because they should know what they're going to get. * Nope, they don't know. * But they can use the similarity to chicken we expect * To show that the technique works * Technique: They build lots and lots of synthetic peptides, build spectra for those peptides, and then use those spectra as the "database" to compare the ostrich spectrum to * Criticism: * No statistical analysis * Possible response (not done in the paper) * Start with DNA (yes, we can get ostrich DNA), sequence it, figure out the protein. * Proof with an independent technique Side note: Collagen is conserved across species. There are minute differences, even between a chicken and a newt. * The other species are used to predict where there are likely to be changes in collagen * How many geneomes can we work with? * There are only about fifteen-twenty full genomes in the database, so we don't get much data to start with. Important lesson from biology: Use different techniques. If you get the same answer from different techniques, you have more confidence. * Note: There's another paper in the same issue of science that uses immunological tests to establish the collagen * Antibodies are things on the outsides of cells that bind to things on the outside of cells * If we can make a flourescent antibody to a protein of interest, we can see whether or not the protein has been expressed Criticisms: * Did not discuss possible problems * The chicken conclusion doesn't really belong Contributions of the paper * Process for identifying peptides that have not previously been MS * "Proof" that proteins can survive in fossils * We can use the process to identify fossil proteins * It's a really wild idea * And we see it in more detail in Schweizer et al. 2009 * "evidence" that some dinosaur collagen closely resembles chicken collagen * Questionable. Not really a conclusion that they were trying to make. More the popular press's approach. Did the scientific method work? * The feedback was useful * But the tone was not good BREAK! What is interdisciplinarity? (at least in this class) beautiful thing that happens when you get 2 people together to make something new Developing something new drawing from more than one discipline Having knowledge in both disciplines having two disciplinarians come together to do new work Can you be interdisciplinary if you aren't yet a disciplinarian Can two relative novices in fields do interdisciplinary work? Yes - it might be beneficial to be less ingrained in a field, so you are willing to be flexible. Other disciplines can contribute - knowledge of statistics; 3 kinds of exercises Wet Lab - doing experiments in the real world CS Lab - writing algorithms Web Lab - using existing tools and data Things you've learned (so far) from doing the Favism lab (Thursday) * There's more data on the typical page than you can deal with * There's a benefit in talking things out from each other