BME100 f2014:Group21 L6

BME 100 Fall 2014	Home People Lab Write-Up 1 | Lab Write-Up 2 | Lab Write-Up 3 Lab Write-Up 4 | Lab Write-Up 5 | Lab Write-Up 6 Course Logistics For Instructors Photos Wiki Editing Help
OUR COMPANY Name: Dustin D. Burley Name: Mackenzie Konves Name: Scott Boege Name: Nicholas Goettl Name: Travis Tibbs Name: Lionel Davis LAB 6 WRITE-UP Bayesian Statistics Overview of the Original Diagnosis System The PCR reactions done in the BME100 lab were designed to mitigate error. Each group tested 2 patients, meaning a total of 68 tests were done (there were 34 groups). Since each patient had their DNA ran three times there was little chance for error (a total of 208 PCRs were run!) Even after this was done there were still things put in place to reduce the risk of error. When it came time to do the image processing there were three ImageJ tests done on each of the DNA samples (3 samples per patient would mean a total of 9 tests were run for one patient) to ensure accuracy in the diagnosis. Not only where the quantity of tests a way to prevent error, but the quality of the ImageJ calibrations served as a check on error. These photos were also taken multiple times to ensure that the images were good enough to use. Since it was so closely calibrated it would be hard to misread the image and give the wrong diagnosis. Since there was so much data in the final spread sheet the Bayesian statistics could be used effectively. A large sample size made determining the actual likelihood of a false positive or negative much easier. What Bayes Statistics Imply about This Diagnostic Approach The Bayesian statistic calculations 1 and 2 conclude that the individual PCR replicates are reliable in detecting whether or not patients have a disease SNP. For both calculation 1 and calculation 2, the Bayes values were close to 1.00 (100%). The high Bayes values indicate that the tests were reliable. Error could have been from the SYBR Green being exposed to light before being used in the fluorimeter, the method of adding primers or transferring liquids, mistakes in the use of the ImageJ software, or a flaw of the Open PCR machine. If the SYBR Green was exposed to light before use in the fluorimeter, the fluorescence of the drop could change. False positives or negatives could occur because of the exposure to the light. Error could occur from the process of transferring primers and enzymes, as the liquids could not be mixed properly and the target gene may not have been copied enough. Human error in calculations using the ImageJ software could lead to either wrong calibration results or wrong values for the patients. A wrong calibration line would skew the results of the two patients even if the patients' drops are analyzed correctly. The Open PCR machine could have had errors, but groups were allowed to leave before the process was complete. There could have been temperature control problems, but the group was not present to identify the issues. The Bayesian statistic calculations 3 and 4 conclude that individual PCR replicates are not reliable in diagnosing or predicting the disease. The Bayes values for calculations 3 and 4 were very small. The small Bayes values indicate that the tests were not reliable. Overall, the Bayes values show that the PCR tests are reliable in detecting a disease but not reliable in predicting a disease. Computer-Aided Design TinkerCAD Using the TinkerCAD ruler tool helped us see how precise designing in a virtual setting can be done efficiently and quickly. The shape tools made it easier to combine small details into one overall product. The shapes also made it easier to put on the computer what you were originally visioning in your mind. The boat that was made was a simple design, but it incorporated many of the features of TinkerCAD. Making the boat in TinkerCAD helped us better design a better version of the PCR machine. Our Design Feature 1: Consumables Kit Feature 2: Hardware - PCR Machine & Fluorimeter