BME100 s2017:Group2 W1030AM L6

BME 100 Spring 2017	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: Anh Nguyen Name: Sheridan Hill Name: Thomas Andrews Name: Kiryl Sheleg Name: Derall Riley Name: Skye Russell Our Brand Name LAB 6 WRITE-UP Bayesian Statistics Overview of the Original Diagnosis System In this BME100 lab, there were seven teams of six students. Each team diagnosed two patients for a total of fourteen patients. To prevent (or reduce) error, we tested each patient three separate times, and we took three different pictures for each of the tests, giving us nine pictures for each patient. We also had a positive and negative control to compare to the patients. Steps to prevent error were also taken in ImageJ. To best calibrate the images, we calculated the levels of SYBR Green I in the droplet, and we found the average number of green pixels in the black background. Doing this, we were able to distinguish the droplet from the background by subtracting the levels of background green from the levels of the droplet's green. Overall, the class had a fairly high success rate with these tests, matching the doctor's diagnoses 11/14 times. The class only had a total of one inconclusive test of the fourteen. Additionally, every group got their results in on time, so there was no blank data. What Bayes Statistics Imply about This Diagnostic Approach When trying to determine if the patient will get a negative final test conclusion, given negative diagnostics signals, the Bayes value determined was over 1.00 (100%). This is a really good result because the accuracy is perfect. However, when looking at a patient’s probability of a positive final test conclusion, given a positive diagnostics signal, the Bayes value determined was close to 0.50 (50%). This value is not desired, as we want a value closer to 1.00. The PCR results are problematic and not reliable because both numbers should be close to 1.00, not just one of the numbers. This essentially means that if it is concluded a person does not have the disease, they know they absolutely do not have it. But not everyone who has the disease SNP will know they have it, which is not good because it will go untreated. The results for predicting if the patient will not develop the result was great. The Bayes value was almost 1.00 (100%), showing that if the test results were negative, we can say the patient will not develop the disease. However, when testing if the patient will develop the disease, the PCR results were not as accurate, coming to roughly 0.7 (70%). If the patient tested positive for developing the disease, they will not necessarily develop it. But if they tested negative for developing the disease, they are almost guaranteed to not develop it. The reliability of PCR was not too bad for predicting the development of the disease. When creating the PCR mix, even though we used a micropipette, the volumes were not exactly what we needed. There could have been some error in the final volume of the solution that we predicted because the micropipette still isn’t perfect. This is a result of human error while using the micropipette as well as device error in the micropipette itself. When taking the pictures for this lab, we used a smartphone camera on a stand set a certain distance from the fluorimeter. This distance was supposed to be constant, but error could have occurred when moving the smartphone camera. The distance was not necessarily exactly the same when we moved the phone back, possibly resulting in inconsistent results. During the ImageJ portion of the lab, the oval used to capture the pixels for the green image may have been too small or too large for some of the images. This would give incorrect measurements of mean gray pixels and negatively affect the final results, which heavily relied on this measurement. Intro to Computer-Aided Design 3D Modeling The software our team used was SolidWorks during the Computer-Aided Design section of the lab. We had some prior experience using SolidWorks in BME182 as well as the first portion of BME100, but we only were able to do some of the more basic functions. Regardless, SolidWorks the go-to software for developing our 3D design because it was the one we knew how to use the best. Then, we were able to create the fluorometer from scratch in a reasonable amount of time. We also were able to learn some new tools within the program. This is because the fluorometer design required us to use tools we were not familiar with. After the completion of the design, we are more comfortable with SolidWorks and are able to use more of its functions than before. Our Design We took the basic design of the fluorometer and added a few components to improve it. We added a camera off to the side of the box that is attached near where the slides would be placed. It sits above the small nook where the slide are placed, so that it will not impede the slides' movements. It is also close enough to take a quality high-res photo. Inside the fluorometer, there will be an ImageJ add-on to automatically test the image. There will be a screen on top of the box (cut-out part) that will be able to display the results of the ImageJ. There is also a USB port to the side, so the results can be transferred to a computer. This will solve some of the problems that we experienced in the lab. The camera will be at a constant distance from the drop, so we do not have to use a phone camera. The ImageJ will be in the fluorometer to keep everything in the same place. The USB port will allow us to still be able to access the information on a computer. Feature 1: Consumables List of consumables: Several strips of empty PCR tubes Disposable pipette tips PCR reaction mix Primer solution SYBR Green I solution Glass slides The Consumables Kit will have the fundamental materials required to conduct PCR and utilize the fluorometer for DNA product analysis. There were no major changes made to the consumables that were used during the lab; our kit is similar. The weaknesses we saw in that consumables kit were not substantial and did not really negatively affect our results. Our kit has the essentials to conduct PCR and does not have too much extra. We thought it would be useful to have the PCR mix and the primer solution for the PCR section of the lab. Glass slides are needed for the fluorometer portion, and SYBR Green I is needed in order to see the fluorescence. The micropipette tips are needed because the tips are changed so often. We also thought PCR tubes are necessary to store solutions and use in the PCR machine. Feature 2: Hardware - PCR Machine & Fluorimeter PCR Machine: The OpenPCR system will be used to amplify the target DNA sequences for analysis using the fluorometer. Once the PCR mix and DNA primer mix are mixed in appropriate quantities and labeled, they will be placed in the OpenPCR thermocycler, which will amplify the specific DNA sequence. This is the exact same process and machine as we used in class. We did not see a big problem with the PCR machine that needed to be addressed. Fluorimeter: The bigger problems we are addressing are with the fluorimeter. The fluorimeter will be used to analyze the amplified PCR products for the presence of specific DNA sequences. DNA mix and SYBR Green I solution will be mixed and a drop of a specific quantity will be placed onto the water-repelling slide. The slide with the specific DNA mix will subsequently be placed into the fluorimeter chamber. Several photos will be taken with the fluorimeter inside a box, and the photos analyzed with ImageJ software. The fluorimeter system will be redesigned based on the 3D modeling of the design that was conducted. Specifically, the new fluorimeter will address the weakness that smartphone and other cameras provide when taking pictures of the drops. Such cameras have varying resolutions and features that may affect the analysis of the samples with ImageJ software. Instead, the camera will be attached to the fluorimeter system. This will increase stability, since other cameras are prone to movement. This will also give control to the distance between each sample to ensure that the contrast is similar in all images. The camera will have certain specifications that aid in analysis. For instance, the ISO, exposure, and saturation will be preset to a very high value while the contrast will be preset to a very low value. Two additional features of the new fluorimeter will be a screen and USB port. The USB port will allow for easy transfer of fluorimeter data onto a computer. The screen will display the results of the ImageJ that the machine will compute. The weaknesses we saw were in human error more so than machine error. To fix this, we enabled the fluorimeter to do more work. Giving it its own camera eliminated what we had to do with the camera (positioning it in the same place every time). The internal ImageJ processing will allow the machine to calculate the fluorescence based off of the pictures it takes. The screen and the USB port allow easier use for the person using the machine.