BME100 f2017:Group12 W1030 L6

From OpenWetWare
Jump to: navigation, search
Owwnotebook icon.png BME 100 Fall 2017 Home
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
Wiki Editing Help
BME494 Asu logo.png

Dunder Mifflin Tempe

Name: Jasmine Barboa
Name: Alivia Ankrum
Name: Stephanie Santiago
Name: Alec McCall
Name: Jake Taylor

Dunder Mifflin Tempe


Bayesian Statistics

Overview of the Original Diagnosis System

BME 100 students were given three replicate DNA samples from two "patients" in order to run a Polymerase Chain Reaction (PCR) in order to detect SNP in the patients. Using pipetting techniques, the DNA as mixed with a primer mix to put into an OpenPCR machine. Next, 8 PCR reaction samples were mixed with the buffers. The samples were mixed with the buffers and then SYBR GREEN solution and a liquid containing double-stranded Calf Thymus DNA (0.25, 0.5, 1, 2, and 5). These samples were then observed using a fluorescence technique. The division of labor throughout the 17 teams and 6 students allowed for the master chart of statistical analysis to be made while each student helped contribute the data and interpretation needed. There were many steps done to eliminate as much error as possible. This was done by using multiple replicates per patient to use as "safety nets" for samples. Along with that, there were multiple PCR controls to observe the samples which allowed the teams to take multiple pictures of the samples for the fluorescence readings, letting students play around with different images within ImageJ calculations for each specific PCR sample. Once all the calculations were done, each group sent them in to make one master statistical analysis sheet which allowed teams to calculate the statistics of the PCR tests and how accurate they were. Overall, it seemed the class had pretty accurate results compared to the doctor's actual diagnoses and there were only three inconclusive results. The experiment seemed to run well amongst the whole class and the only problems that were encountered seemed to be the confusion on calculating the results of the ImageJ calculations. However, after team collaboration, the correct method for calculations and graphing was determined and carried out.

What Bayes Statistics Imply about This Diagnostic Approach

The result for calculation 1 shows the probability that the sample actually has the cancer DNA sequence, given a positive diagnostic signal. Our result was close to 75% accuracy, which means that each individual PCR replicate is relatively reliable in concluding the final conclusion of whether a patient has that disease SNP or not, as it would be accurate about 3 out of 4 times.

Calculation 2 determined the probability for a patient to receive a negative test result given there was a negative diagnostic signal. It also described the specificity of the system regarding the ability to predict the disease. The results for calculation 2 were close to 100% meaning that there was reliability in the ability of the PCR reaction to detect negative test conclusions for SNP.

The results for calculation 3 was the frequency of cancer "yes" diagnoses was close to 50% and the frequency of positive DNA test conclusions was a little more than 50%. The frequency of positive given cancer yes was a little more than 60%, so the frequency of cancer "yes" was positive were a little more than 50%.

The results for calculations 4 were used in order to the determine a negative cancer finding and a diagnosis of "no" from a doctor. The reporting of a negative cancer finding was found to be slightly below 50%, while the diagnosis of doctor saying "no" was determined to be marginally above 50%. Overall, the probability of an overall negative finding and diagnosis was calculated to be a solid 100%.

One source of human error that could have occurred during the PCR and detection steps was the focus of the camera portraying the green intensity for the PCR reaction samples inaccurately, which would have resulted in the wrong detection for whether it was positive or negative. Another source of error could have been the incorrect area selection in ImageJ, by adding some of the irrelevant colors of the drop, which would have altered the the RAWINTDEN density of the drop. Lastly, the incorrect dilution of the PCR reaction sample would have added a more intense or less intense green appearance that would have also skewed the appearance of the drop, and thus, the RAWINTDEN for each sample.

Intro to Computer-Aided Design

3D Modeling
The software that we used was Solidworks and that program is really useful for making projects like this. It is hard to learn, but when you get the hang of it it is super easy. Only one group member really has had any past experience with Solidworks. This project tested the limits to try and create something new for the lab set up because it is hard to come up with something to improve on. Solidworks is a great program if you know what you are doing and how you want to do.

Our Design

Solidworksboxthingy1.jpeg Solidworksboxthingy2.jpeg Solidworksboxthingy3.jpeg

This design is for helping take better pictures of the cyber green mixtures to determine what we are finding in our research. The design has a drawer for the flourimeter. The drawer will have batteries in it so that it can power the laser. Then the drawer is placed in the compartment where it will sit and when it is inside then the drop is placed on the glass between the lasers. Then the lid will be closed so that it is fully dark inside. There is a little hole for the camera on any phone will take a picture from. There is an adjustable stand for the phone to sit on so that it can properly be focused and still. The distance is set up so that most phones will focus at its best. When everything is ready then everything must be closed and the pictures can be taken.

Feature 1: Consumables

Our consumable kit product list is as follows:

  • PCR Mix
  • Primer Solutions
  • SYBR Green Solution
  • Buffer
  • Glass Slides
  • Plastic tubes (standard size but company provided)
  • Pipettor (standard design but company provided)
  • Pipettor tips (not standard design and only provided by the company)

The pipettor tips have been redesigned because the original design allows the tips to frequently fall off of the pipettor. This would cause PCR sample mix and primer solutions to be wasted due to the loss of the pipettor tip. Our company design prevents this problem by including a "bump" on the tip to act as a lock when connected to the pipettor.

Feature 2: Hardware - PCR Machine & Fluorimeter

We didn't make any changes to the OpenPCR machine because we thought it ran satisfactory. The only thing that we improved on was the fluorimeter system because we had trouble trying to get good pictures for the project, specifically, trying to keep it dark to take the picture. So, we improved on this by adding a separate component that allows for almost complete darkness while imaging the PCR results using the SYBR Green I solution. Increasing the darkness of the environment allows for a better image to be taken of the result, which will thus allow for more accurate data analysis and results. A separate box is to come with the PCR machine, to achieve these high quality images. This box includes a drawer, that the user can easily transfer their sample from the PCR machine to, and once closed, it is completely enclosed in darkness, aside from a tiny hole in the side of the box. This hole in the side of the box creates the desired user interface, allowing the user to put their phone up to the hole and take a picture of their result in almost complete darkness. Of course, the lasers activating the SYBR Green solution will be present in the sides of the box. Lastly, there is a built in and adjustable stand for the user's electronic device used to take pictures. That way, the user does not have to hold their phone and risk shaky hands affecting the outcome of their image, and thus, their results.