BME100 f2014:Group3 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: Marissa Seelhammer Name: Brianna Denuit Name: Farhad Eghbalian Name: Shane Mitchell Name: Catherine Piatak Name: Nivenka Mahesh LAB 6 WRITE-UP Bayesian Statistics Overview of the Original Diagnosis System In order to test patients for the disease-associated SNP, DNA samples are amplified through polymerase chain reaction in an Open PCR machine and analyzed using imaging software. The BME 100 lab class, divided into 34 teams, analyzed the DNA of 68 different patients and determined whether their results for the disease were positive, negative, or inconclusive. Each team was given exactly two patients’ DNA samples and tested them against positive and negative controls to determine a diagnosis. In order to prevent error, each patient had three different samples of DNA, all of which were to be tested separately so that the chance of a misdiagnosis was minimized. Additionally, ImageJ, the software used to analyze the pictures of the amplified DNA drops, was specifically calibrated to measure the most accurate amount of green color (SYBR Green I) in the pictures while filtering out the blue and red light. Error was also minimized by having the groups analyze three separate drops for each patient using ImageJ and averaging the results, ideally making the data more reliable. Out of the 186 total patient DNA analyses, there were 92 positive diagnoses and 73 negative diagnoses, leaving 21 tests inconclusive. When the three trials of each patient were averaged together, the final outcome was 30 positive, 24 negative, and 6 inconclusive. There were also 6 patients whose data was never submitted. Out of those with a positive diagnosis, 89% actually had the positive DNA, and out of those with a negative diagnosis, 77% actually did not have the positive DNA. This indicates that the process used does successfully diagnose the patient, however, ideally the percentage of a correct diagnosis would be above 95%. What Bayes Statistics Imply about This Diagnostic Approach Calculation 1: This shows how accurate the test is to detect a positive result for SNP (sensitivity). The result for this test showed that the probability of a patent getting a positive result for having SNP from the PCR machine given that they have the disease is 89%. Calculation 2: This shows how accurate the test is to detect a negative result for SNP (specificity). The result for this test showed that the probability of a patent getting a negative result for having SNP from the PCR machine given that they do not have the disease is 77%. In other words, both of these show accuracy but also have room for improvement. This room for improvement comes from the error that occurred during the lab. These were a result from both human error and device error. Human errors include error in the preparation for the test. Specific tests that were shown to be positive should have displayed a negative result. This could be as a result from mixing up the samples. In the case for our group, the negative control (in other words, the sample that we knew should have displayed a negative result for the disease) displayed a positive result for SNP. The mix up in the samples probably occurred before we actually deceived the samples. This error could have also been a result of improper disposal of pipet tips before and after each use during preparation, thus causing a mix of samples. Device error includes not running the same amount of cycle in the open PCR machine. Each group's machine ran a different amount of cycles and some machines simply did not work at all. Receiving a positive result for our patient sample when it should have been negative left error in the Bayes values. Calculation 3: This shows how accurate the PCR machine is to predict the disease in a patient that would develop the disease. The test was not accurate given that it had a 43% efficiency rate. Calculation 4:This shows how accurate the PCR machine is to predict that a patient would not develop the disease. The test was even mess reliable than calculation 3 because it only showed a 27% efficiency rate. Computer-Aided Design TinkerCAD The TinkerCAD tool is used just like the software SolidWorks, it is there to create 2-dimnsional or 3-dimensional figures online. TinkerCAD was used to make our 3-dimensional PCR. First we made all of the different pieces of the PCR separately and it was 2-dimensional. Then after all the separate pieces were made, it was put all together to make the final, 3-dimensional, product. Our Design Feature 1: Consumables Kit The kit will be packaged with the new and improved color coded plastic vials to replace the bland clear ones. Color coded plastic vial, will add a significant strength to the overall procedure process. Each color coded plastic tube (blue, red, green, black and clear) specifically will be used for designated substance. This major improvement considerably reduces the error upon filling the tubes, enhance operator confidence (in regard to procedure steps), speed the process of filing the tubes, reduce the cost of operation (substance, tips, vile and the time that must be use to redo the process), create a systematic approach to the procedure. Furthermore, the black vial may specifically be used for SYBR green substance. The black tube, greatly reduce the light exposure to SYBR green substance, increase the SYBR Green sensitivity, reduce the cost of over use of SYBR Green, eliminate the steps of aluminum foil covering and increase the speed of procedure. Feature 2: Hardware - PCR Machine & Fluorimeter