BME100 f2014:Group31 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
LAB 6 WRITE-UPBayesian StatisticsOverview of the Original Diagnosis System From the DNA test of a disease-associated SNP, a Bayesian calculations were done in order to determine probability of an accurate diagnosis. The division of labor was 34 teams averaging around 6 students each that diagnosed 68 total patients. Each patient was tested for a positive or negative reading, which was then compared to a positive or negative control. In this way, accuracy was tested. In order to further prevent errors, 3 trials per patient were run, to ensure proper reading. Two PCR controls, one positive with DNA in it and one negative without any DNA, were utilized to provide a benchmark for ImageJ pixel comparisons. Within ImageJ itself, the pixels were calibrated to match the drop size used in the measurement. Finally, three pictures of each trial were taken in order to give a more accurate reading of the drop. The following Bayesian calculations were done using the results shown below. The equation used to calculate probability of reading accuracy was: P(A|B)=(P(B|A)*P(A))/P(B) P(A|B) is the Probability of A given B P(B|A) is the Probability of B given A
Probability of positive final test conclusion given a positive PCR reaction A=.48 B=.49 P(B|A)=.89 P(A|B)=.87 Probability of negative final test conclusion given a negative PCR reaction A=.39 B=.39 P(B|A)=.77 P(A|B)=.77
Probability patient will develop disease w/ positive final test conclusion A=.34 B=.48 P(B|A)=.43 P(A|B)=.30 Probability that a patient will not develop disease w/ negative final test conclusion A=.66 B=.39 P(B|A)=.27 P(A|B)=.46 What Bayes Statistics Imply about This Diagnostic Approach
Although the measurements produced Bayes measurements close to 1, there were potential sources of error that detracted from their accurate measurement. One potential error may have arisen because of the placement of the camera. The slide may not have been entirely level with the camera, producing an image with a skewed view of the drop. This may have caused the fluorescence to be under or overestimated to the point that the SNP is not detected. Another source of error may have come from the actual structure of the box. Had there been crevices within the construction, light may have been increased, causing the fluorescence to be inaccurately measured. This would have possibly created a positive diagnosis for the SNP when there was no SNP sequence. Finally, the image quality itself is a potential source of error. With differing camera phones used by each group, the image quality may have varied. In addition, even though each camera was placed a set distance of about 4 cm. away from the slide, uncertainty may have been caused from differing levels of zoom, which would increase the image of the drop but alter the fluorescence measurement. All these errors may have resulted in the faulty detection of the SNP within the DNA. The calculations for 3 and 4 are considered the "diagnosis." These Bayesian calculations directly find the probability and accuracy of the device in predicting contraction of the disease. These are the sensitivity and specificity in predicting the disease. Both these values were both beneath .50 and thus, considered very small. This indicates a lower degree of accuracy in predicting the disease, especially when compared with detection of the SNP in calculations 1 and 2. These small values that result from the calculations are indicative of the necessity for redesign should a device be aimed towards predicting the disease.
TinkerCAD
Works Cited: This image was directly copied from KickStarter.com and can open the link below! https://www.kickstarter.com/projects/930368578/openpcr-open-source-biotech-on-your-desktop
TinkerCAD
Feature 2: Hardware - PCR Machine & Fluorimeter | |||||||
