Basic Statistics; Standards iScientic Communities I!Module 3, Lecture 3!20.109 Spring 2010!n2!Lecture 2 review!•! What properties of hydrogels are advantageous for soft TE?"•! What is meant by bioactivity and how can it be introduced?"•! What are the two major matrix components of cartilage and how do they support tissue function?"3!Topics for Lecture 3!•! Module 3 so far, and Day 3 plan"•! Introduction to statistics"–! condence intervals"–! t-test"•! Standards in scientic communities"–! general engineering principles"–! standards in synthetic biology"–! standards in data sharing"4!Module progress: week 1!•! Day 1: culture design"–! What did you test?"•! Day 2: culture initiation"–! Cells receiving fresh media every 2 days"© BD. All rights reserved. This content is excluded from our Creative Commons license. For more information, see http://ocw.mit.edu/fairuse.5Module day 3: test cell viability!Green stain: SYTO10 = viability! Assay readout: Red stain: ethidium = cytotoxicity! uorescence!Working principle?! Relative cell-permeability!!Statistics review: basics!•" Essential concepts: standard deviation (s), mean ( ), sample size n, degrees of freedom DOF!•" Normal (Gaussian) distribution!x-axis:!y-axis:!1 s includes! ______ %!of the data!Condence intervals (CI): principle!•" = 60 (sample/measured mean) !•" 95% CI calculated to be ± 3!•" Thus: 95% likely that the range 60 ± 3 contains the population (true) mean µ!–" exact denition is subtle!•" 90% CI: µ= ± a where a < 3 a > 3 a = 3 ?!•" Consider betting example!•" What about n?!Calculating condence intervals (CI)!•" t is tabulated by DOF vs CI–" DOF = n - 1 !•" In Excel, us TINV function!–" input p-value = (100-CI)/%!100!Introduction to t-test!•" Every statistical test!–" has assumptions!–" asks a specic question!–" requires human interpretation!•" Some t-test assumptions !–" normal distribution (cf. Mann-Whitney test)!–" equal variances (type 2 in Excel; type 3 unequal)•" Question!!Calculating t-test signicance!•" If tcalc > ttable difference is signicant !•" In Excel, us TTEST function!•" Excel returns p-value ! condence level (CL)!•" 1-tailed vs. 2-tailed test!DOF!ttable!Assignment for report!•" Get live cell count and/or live cell percent values for both culture conditions!•" Calculate 95% CI for both means!•" Plot means on bar graph with CI error bars!•" Apply t-test to the means!–" For multiple comparisons, ANOVA is better!–" Comparing many means requires correction!–" Remember, p = 0.05 means 1 in 20 false positives!!Interlude: intersection of science and commerce!1.! HeLa cells "http://www.colbertnation.com/the-colbert-report-videos/267542/march-16-2010/rebecca-skloot ( ~00:30-3:00 )"2. Patenting genes" “Judge invalidates human gene patent” NY Times March 2010" “Metastasizing patent claims on BRCA1” Genomics May 2010 "13!Thinking critically about module goals!•! Purpose of experiment"–! Local"–! Global"•! All well and good, but…"•! Can we move beyond empiricism – tissue engineering"•! E.g., broadly useful biomaterials"–! goal: control degradability over wide range "–! “a lot of chemical calculations later, we estimated that the anhydride bond would be the right one”"–! Robert Langer, MRS Bulletin 31(2006)."14!Engineering principles, after D. Endy!•! D. Endy, Nature 438:449 (2005)"•! Is biology too complex to engineer, or does it simply require key “foundational technologies”?•! Systematic vs. ad hoc approach"•! Abstraction"–! software function libraries"–! copy-editor vs. editor"•! Decoupling"–! architecture vs. construction"–! design vs. fabrication"•! Standardization"–! screw threads, train tracks, internet protocols"–! what would we standardize to engineer biology?" "Public domain image Wikimedia Commons)(!15!Application to synthetic biology!""See D. Endy, Nature 433: 449!•! D. Endy, Nature 438:449 (2005)"•! Synthetic biology, in brief: “programming” cells/DNA to perform desired tasks"–! artimisinin synthesis in bacteria"–! genetic circuits "•! Abstraction"–! DNA ! parts ! devices ! systems"–! materials processing to avoid unruly structur•! Decoupling "–! DNA design vs. fabrication (rapid, large-scale•! Standardization"–! Registry of Standard Biological Parts"–! standard junctions, off-the-shelf RBS, etc. "es)'Can I have three inverters?''I need a few DNA binding proteins.''Get me this DNA.'PartsDevicesSystemsDNAImage by MIT OpenCourseWare.16!Data standards: what and why?!•! Brooksbank & Quackenbush, OMICS, 10:•! High-throughput methods are data-rich"•! Standards for collection and/or sharing!•! Reasons"–! shared language (human and computer)"–! compare experiments across labs"–! avoid reinventing the wheel"–! integration of information across levels"•! Examples"–! MIAME for microarrays"–! Gene Ontology (protein functions)"•! Who drives standards?"–! scientists, funding agencies, journals, indu94 (2006)"stry"http://www.geneontology.org/Screenshot image captured April 2010. Courtesy of the Gene Ontology. !Used with permission.17!Lecture 3: conclusions!•! Condence intervals and t-tests are two useful statistical concepts."•! Standardizing data sharing and collectiois of interest in several BE disciplines."n Next time: discussion of standards in TE; more about cell viability and microscopy!© Registry of Standard Biological Parts. License CC BY-SA.Microarray data!See: D. Endy, Nature 438:449 (standardized biological “parts”)!This content is excluded from our Creative Commons license.For more information, see http://ocw.mit.edu/fairuse.MIT OpenCourseWarehttp://ocw.mit.edu 20.109 Laboratory Fundamentals in Biological Engineering Spring 2010 For information about citing these materials or our Terms of Use, visit: