BME100 f2015:Group17 1030amL6

From OpenWetWare

Jump to: navigation, search
BME 100 Fall 2015 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
Image:BME494_Asu_logo.png


Contents

OUR COMPANY

Andrew Nelson: studentRole(s)
Andrew Nelson: student
Role(s)
Augustas Laurin
Augustas Laurin
Eric Slovak: studentRole(s)
Eric Slovak: student
Role(s)
Name: Taylor Underwood
Name: Taylor Underwood
Name: Julianna Acero
Name: Julianna Acero
Name: student
Name: student


LAB 6 WRITE-UP

Bayesian Statistics

Overview of the Original Diagnosis System

Our group of five was divided between the PCR sampling and micro pipetting, and performing the ImageJ portion of the lab. The two group members assigned to prepare the solutions for PCR acquired all supplies needed for the wet lab. This included micro pipet test tubes, test rack, extracted DNA, two primers, fluorimeter and light shield. The same two people, wearing proper wet lab attire, performed all micro pipetting creating samples ready for the fluorimeter. Then the mixed solutions that were just crested were transferred to the fluorimeter plate, which had hydrophilic holes surrounded by hydrophobic space forming a large drop of solution allowing the fluorimeter test to be performed. Once the light shield was placed, three drop images were taken in an attempt to reduce error due to blurred pictures so the ImageJ could successfully run the test. Our individual group diagnosed 2 patients and our class overall diagnosed 34 patients. The other three members of our team conducted the ImageJ portion of our lab by taking our fluorimeter pictures and calculating MEAN (of RAWINTDEN DROP - BACKGROUND), PCR Product Concentration (μL/mL), Total Dilution, and Initial PCR Product Concentration (μL/mL). Then we compared these values to the Calf Thymus calibration curve performed earlier. From the calibration curve we found how much DNA was produced by the polymerase chain reaction (PCR). These results where then collected from all 17 groups in class and shared in order to analyze the data using Bayesian statistics. Of the 34 final patient results, 4 came back inconclusive or no data, 13 came back positive, and 19 came back negative.

Many steps in this lab were designed to prevent error. We were given three replicates per patient allowing for three independent tests to be run resulting in data with overall less error. If we would have just had one replicate, our data would be more prone to error due to less sampling points. We also had PCR positive and negative controls to compare to our PCR results. This reduced error be allowing our group and ImageJ to have reference points to compare all other points. We also ran a Calf Thymus calibration curve and an ImageJ calibration to ensure precise data. Finally, we took three pictures of each drop sample totaling 18 for the Calf Thymus calibration curve and 24 for the PCR. Taking three pictures allowed a mean to be calculated ensuring more precise data than a single picture.


What Bayes Statistics Imply about This Diagnostic Approach


Calculation 1 determined that the probability of a positive final test conclusion was about 50% (41%), the probability of a positive PCR reaction was about 50% (45%), the probability of a positive PCR reaction given a positive final test conclusion was about 100% (85%), and the probability of a positive final test conclusion given a positive PCR reaction was about 75% (77%). With a probability of about 75%, our test was good at successfully detecting a true positive result as 100% would be perfect detection.

Calculation 2 determined that the probability of a negative final test conclusion was about 50% (53%), the probability of a negative PCR reaction was about 50% (49%), the probability of a negative PCR reaction given a negative final test conclusion was about 75% (82%), and the probability of a negative final test conclusion given a negative PCR reaction was about 100% (89%). With a probability of about 100%, our test was really good at successfully detecting a true negative result.

Calculation 3 determined that the probability that a patient will develop the disease was about 25% (31%), the probability of a positive final test conclusion was about 50% (41%), the probability of a positive final test conclusion given that a patient will develop the disease was about 25% (31%), and the probability that a patient will develop the disease given a positive final test conclusion was about 25% (24%). With a probability of about 25%, our test was not very good at having a positive final test conclusion predict the development of the disease.

Calculation 4 determined that the probability that a patient will not develop the disease was about 75% (69%), the probability of a negative final test conclusion was about 50% (53%), the probability of a negative final test conclusion given that a patient will not develop the disease was about 75% (71%), and the probability that a patient will not develop the disease given a negative final test conclusion was 100% (92%). With a probability of about 100%, our test was very good at having a negative test predict the disease will not develop.

Sources of error effecting the Bayesian statistics would be flawed fluorimeter and ImageJ testing. Through ImageJ we determined our results. If the procedure of the lab was not followed exactly, if our samples were contaminated, if we left the cyber green in the light too long, if our picture quality was bad, or if our lighting was bad, it could all contribute to error. Also, with only 34 data points, any changes in outcome will substantially effect our probabilities so a larger sample size would reduce error. Finally, if the PCR machine didn't function correctly by not reaching the temperatures desired in the cycle, then our data would be flawed for our whole batch making error hard to detect.



Intro to Computer-Aided Design

TinkerCAD
TinkerCAD is a free online tool that allows users to create 3D models of anything they would like. In this project, we used it to look at a 3D model of the open PCR system and design an improvement for the system. While we don't have anything to compare it to, TinkerCAD was fairly user friendly; it didn't take too long for our team to begin using it. That being said, TinkerCAD is a great tool for anybody interested in any kind of 3D modeling.

Our Design

Our Design
Our Design


While looking at the original 3D model of the open PCR system, we noticed that the cooling block looked almost identical to a CPU cooler as used in desktop computers. We quickly realized that we could make the cooling process a lot faster if we instead used a water cooling loop. Water cooling loops are often used in high end gaming PC's where the CPU is overclocked therefore requiring much more efficient cooling to maintain temperatures at an operable level. An example of such a cooler is the Corsair H100i. Our design features a cooling block that would attach where the old cooling blocked attached connected via tubes to a radiator that would cool the water flowing through the system. While this is a more expensive option, it would greatly improve the efficiency of the cooling and reduce the time required to wait while cooling.


Feature 1: Consumables

Our redesigned PCR will use the same consumables as the original used in class. Our company would make a minor change to have the respective PCR mix, primer solution, SYBR Green solution and buffer all be in pipet tips connected together for easier compartmentalization to reduce the chance of mixing wrong.

Strength: Consumables are cheap, the micro pipette is precise, plastics are disposable, and the tubes fit well inside the PCR machine Weakness: Cross-contamination, wastes a lot of pipet tips

Feature 2: Hardware - PCR Machine & Fluorimeter

OPEN PCR

Strengths: Provides accurate measurements, easily tested samples, easily accessible

Weaknesses: time consuming, display was small

We identified "time consuming" as a major weakness of the current PCR. Our new PCR has improved heating and cooling system that decreases the time need for each thermo-cycle thus reducing the total time needed for the PCR reaction to run. The size of our PCR machine will be larger limiting it for lab use only but the reduced run time will be suitable for a lab setting.

FLORIMETER

Strengths: hydrophobic/hydrophilic plate worked well, laser was positioned at an appropriate level to shine through the solution drop

Weaknesses: human error prone camera set up, slides were hard to slide into the stand, shade device

If our group would have redesigned the fluorimeter set up we would have made one minor changes. It would be to have a camera stand integrated into the device for more accurate pictures.





Personal tools