|
==
Easy PCR
LAB 6 WRITE-UP
Bayesian Statistics
Overview of the Original Diagnosis System
The original Diagnosis System that was used to diagnose the patients included two parts, the first part included a color fluorimeter and a second optional part was a gel electrophoresis lab. The first part of the lab was done by ten different groups of six people, however the division of labor for each group was different. For example, in our group two students worked on the fluorimeter images, setting up the tubes of DNA, diluting the solutions and taking the images. The other four students analyzed a portion of the images in the ImageJ software. After the lab each student analyzed all of the images through the use of ImageJ software. This software separated the colored images in order to get the data of the mean color penetration through each sample of DNA. Once the DNA samples were created and analyzed the Image J software was used to compare this data to the calibration data. The calibration curve data was used to find out the concentration of the DNA produced in each PCR sample. Three different drop images were used for each ImageJ calculation to obtain a level of consistency for each sample. We used the same cell phone for each picture of the drops. The PCR samples contained a negative and positive control that was used to compare the tested dros of the calf thymus DNA. The final class data concluded that six people tested positive to have the disease, twelve people tested negative and two trials produced inconclusive results. The challenges our group encountered during this experiment included having to carefully place the drop on the slide, then focusing the camera in a dark environment and then sucking all inklings of the drop back into the micropipette. This was a time-consuming process that required careful attention, which caused us to rush the last few trials, as the class period was coming to an end. This could have affected the data because we were not as careful with the placement and clean-up of the drops, so contamination of samples was possible.
What Bayes Statistics Imply about This Diagnostic Approach
Calculations 1 and 2 (shown below) determined the relationship between each PCR result and the overall conclusions of the tests. Both calculations resulted in a probability of less than 1. This means that there is a relatively high but not perfect probability that a positive or negative PCR result will end with a positive or negative conclusion. These results show the importance of multiple trials as you cannot rely on one trial alone to determine the patient's probability of developing the disease.
Calculations 3 and 4 (shown below) show the probability that a positive or negative result will end with the patient developing or not developing the disease. The probability for developing the disease after getting a positive conclusion was very close to 1 and so a positive conclusion could likely be trusted by a patient. The values are not perfect but they are extremely close to being correct for every test. For the negative conclusion the probability that the patient will not develop the disease is lower than 1 making it less reliable. If a patient got a negative conclusion on their test it is possible that they could still develop the disease due to the lowered reliability of the test, thereby I would recommend patients get additional testing if they were concerned.
There are several sources of error that could have been an issue in this lab. For the first source of human error the phone that was used in our group to take the images could not fit into the stand and thereby it was moved after every trial. This made it so the images were not uniform and the possible differences in distance from the droplet may have altered the results that came from Image J. There were also several systematic errors stemming from the lab set up. First of all the droplets were placed in the same area as the previous droplets and this setup could lead to contamination and problematic results. Additionally, the camera on the phone itself had issues focusing on the droplet so the images came out blurry and led to results that were not precise. All of these errors and more could have led to the lowered probability of the test providing accurate conclusions.
Variable Description Numerical value
A Probability Positive conclusion 0.3
B Given Positive PCR reaction 0.32
P(B A) Probability of B given A 0.79
P(A B) Probability of A given B 0.75
Variable Description Numerical value
A Probability Negative conclusion 0.6
B Negative PCR reaction 0.65
P(B A) Probability of B given A 0.9
P(A B) Probability of A given B 0.83
Variable Description Numerical value
A Probability patient develops disease 0.4
B Positive test result 0.3
P(B A) Probability of B given A 0.67
P(A B) Probability of A given B 0.89
Variable Description Numerical value
A Probability patient will not develop disease 0.6
B Negative test result 0.6
P(B A) Probability of B given A 0.75
P(A B) Probability of A given B 0.75
Intro to Computer-Aided Design
3D Modeling
Our group utilized TinkerCAD to design a possible PCR machine. TinkerCAD was simpler and more intuitive to use than SolidWorks. TinkerCAD has many pre-loaded shapes that made it very quick and easy to cobble together a design for a PCR machine. Since the PCR machine is not filled with complicated structures, we decided that TinkerCAD would be more productive than attempting to design all the primary pieces in SolidWorks. TinkerCAD also didn't require any prior experience to successfully maneuver the different features to efficiently design a PCR machine within the time constraints of the lab. Overall, TinkerCAD was a helpful piece of software that didn't have a steep learning curve, which allowed us to design a professionally-looking PCR machine.
Our Design
Our design is very similar to the original OpenPCR design, it still contains the cardboard box with the phone stand and the trays for the DNA droplet placement, but we designed our PCR machine around the theme of improving efficiency of testing drops of DNA. The major differences lie in the phone stand and the DNA trays. The DNA tray will slide sideways into the cardboard box so that the box does not have to be removed each time a new sample is tested. In our design, the drops can be loaded all at once on the tray and then slid in, one-by-one, through the slits in the side of the cardboard box. We also designed a new phone holder that is made from sturdier material than plastic, either silicone or rubber, something that provides a high amount of friction, to hold the phone firm in place. The holder will also have adjustable straps to hold different sizes of phones.
Feature 1: Consumables
Our product is based on efficiency for large labs to get the results quickly. This means that the cost component of including only the very important products will not be an issue. Our product will therefore include all the products necessary to run the test. This includes the PCR mix, primer solution, plastic tubes, glass slides, micro pipette tools of various sizes, labeled micro pipette tips, and light stand and box to block out light. This product will come with extensive instructions on how to use each tool that is included and then how to run the tests themselves. There are some changes that will be made to the Low Cost Distributed PCR design. First of all, the glass slides that the droplets are placed on will be altered to accommodate several drops and several trials at the same time. The box will have openings on either side to allow the longer glass slide to hang out over one side and then be pushed into place without removing the box lid or moving any of the equipment from their positions. The stand that the phone will sit on will also be altered to allow many sizes of phone to fit into the fluorimeter system. The stand will be fitted with straps to stabilize the phone while keeping it in position to take the images. These straps will likely be made with velcro or elastic to stretch over the phone or adjust to various sizes.
The one weakness that our product will tackle regarding consumable products is in the micro pipette tips. These tips must be removed and replaced with every trial and if a lab is testing large quantities of DNA at the same time they will need to be positive that there will be no risk of contamination. By labeling the tips that will be used the researchers will have a much easier time of determining which tips have already been used and they will have no trouble working with enough DNA to provide a conclusive result.
Feature 2: Hardware - PCR Machine & Fluorimeter
Our system will use relatively the same fluorimeter setup used in the original PCR experiment during the BME 100 Lab. We will keep the general fluorescent light, phone holder, DNA tray and overhead cardboard box. The only alteration will be the slits in the side of the box to allow the DNA tray to be slid in from the side of the box. The new type of hardware our design will implement is the altered box with slits in the sides.
Error creating thumbnail: Unable to save thumbnail to destination
Error creating thumbnail: Unable to save thumbnail to destination
Error creating thumbnail: Unable to save thumbnail to destination
| 
BME 100 Spring 2017
