BME100 f2014:Group32 L6

From OpenWetWare
Jump to: navigation, search
Owwnotebook icon.png 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
BME494 Asu logo.png


OUR COMPANY

Name: Blake Bosold
Name: Royal Boggs
Name: Isaiah Gonzales
Name: Fahhad Ashour
Name: Nava Nozari
Name: Neal Viswanath


LAB 6 WRITE-UP

Bayesian Statistics

Overview of the Original Diagnosis System


The lab experiment consisted of 34 different groups with 6 different people performing PCR and tests on the different DNA samples in order to analyze them; the groups had 3 different samples from 2 different patients and needed to determine whether or not the DNA samples ere positive or negative. The groups in the class ran a total of 186 PCR tests and diagnosed 68 diferent patients with the SNP disease positivity or negativity. There were 3 different samples in order to decrease the error in the experiment and to combat any experimental error or extraneous variables that could affect a single sample and not the others. Each sample was tested using SYBR Green and a fluorimeter which tested for positivity in the DNA samples had they glowed green. For each drop of the DNA sample 80 microliters of the SYBR green solution was mixed in and then images were taken then analyzed in ImageJ to isolate the green within the pictures. The pixels of the DNA were then analyzed and a numerical value was given to them in order to calculate the concentration and to create a calibration curve of the patient samples in order to analyze the DNA. The fluorimeter was a large dark box that isolated light from coming into the box in order to get the highest resolution photos. In the lab the settings of the iPhone camera were also changed so that the highest quality pictures for the setting could be taken. ImageJ is a reliable software that analyzed the images correctly; 3 images per sample were taken in order to reduce the error and uncertainty in the experiment. The amount of trials between all the groups combined also reduced the uncertainty of the entire experiment and the results as well by increasing the sample size of the experiment, thus giving the data a close to normal distribution in terms of error. The split up of labor between different groups lessened the amount of experimental error that could have happened with the magnitude of samples but increased the likelihood of human error happening.

These results were then analyzed using Bayes statistics to calculate the probablity of the results. The final data resulted in 23 positives and 45 negatives however, the data that was given only required 30 positive samples and 24 negative samples. Some of the data was either discarded or blank because groups made errors within the experimental method that biased the results.


What Bayes Statistics Imply about This Diagnostic Approach

The Bayes Statistics approach to this implied that for calculations 1 and 2 the results seemed to be reliable in determine whether or not the patients had the disease SNP or whether they did not have the diseased SNP. The calculations for both were close enough to 1 to be considered alright, but the discrepancies in the data accounted for the large error away from 1. The calculations from 1/2 compared the diagnosis from the fluorimeter and the PCR machine as positive or negative, gave a numerical value for how accurate the PCR readings were. Because these numbers were fairly close to 1 within the error, one can say that the diagnosis method used was fairly accurate.

However the discrepancy in the calculations cannot be ignored. The discrepancies were because of idealizations inherent in the theory as well as errors and uncertainties in the experimental method. The camera placement each time in the experiment might have been slightly off more and more each time, thus propagating the error of the experiment exponentially. The box itself was not closed properly during all pictures as well which could explain the discrepancies in the data; light let into the box distorted the images and made the data processing error ridden. The images taken by the phone itself were neither of the highest nor clearest qualities, thus the images themselves had slight distortion by the camera as well. Error within the experiment could have occurred in the human portion as well; during the experiment the samples could have gotten contaminated by oils or foreign material, different materials not supposed to be combined could have been combined or incorrect amounts of material between the tubes could have been placed in the PCR machine causing unsatisfactory results. The experimental method could be wrong as well; not enough sample could have been used or maybe there was not enough SYBR green solution per droplet in order to test for DNA sampling. Incorrect analysis of the results leading to an incorrect calibration curve could have led to incorrect results as well and thus a less reliable method.

Calculations 3 and 4 give a direct numerical accuracy of the device used in this diagnostic method for the detection of the disease. The values were not statistically significant because of how small they were and thus one cannot say that this device is extremely accurate at measuring the disease based on the data taken. Because these values were not as good as the first two calculations, this suggests that this experimental method design is wrong or needs significant improvement in order to raise the standards of this device and experimental data. Both of the values calculated for the test were very small and below 0.5 which means that there was less than a 50% chance of detecting the correct diagnosis given the positive or negative test.

BME 100 Lab 6 Group 32 table 1.png

1)The calculation of the probability that a patient will get a positive test (SNP), if he/she were given a positive PCR reaction is close to one. Therefore, the PCR can be considered reliable for these calculations within the limits of accuracy. Possible reasons for error:

  • Ratio of the concentration of patients DNA and other reagents was not stable.
  • The volume of reagents in test tubes might have differed in each trial.
  • Light reacting with the SYBR green because of the incomplete closure and blocking out of outside light while taking pictures of the droplets.
  • The PCR machine takes longer time to cool down might disrupt or deform some enzymes or mess calculations.

2)The probability of a patient getting a negative test (SNP) given a negative diagnostic signal is less than the first value yet close to one within limits of accuracy.

Possible reasons for error in values:

  • Ratio of the concentration of patients DNA and other reagents was not stable.
  • The volume of reagents in test tubes might have differed in each trial.
  • Light reacting with the SYBR green because of the incomplete closure and blocking out of outside light while taking pictures of the droplets.
  • The PCR machine takes longer time to cool down might disrupt or deform some enzymes or mess calculations.


3)The probability that a patient will develop a disease given a positive test conclusion is relatively low compared to what was expected (<50%). therefore it proves the unreliability of a PCR in testing for SNP disease.


4)The probability of a patient developing a disease, given a negative final test conclusion, is close to 50%, which is not close to one so the PCR is considered unreliable.


Example.jpg

Computer-Aided Design

TinkerCAD

TinkerCAD was an online software tool used in order to construct different objects using the preprogrammed objects built into the software itself and combining them into different shapes. TinkerCAD first had a detailed summary and lessons on instruction on the process that one can use in order to design different objects. TinkerCAD was used to build an Open PCR machine using objects imported into the software and then assembling them in the CAD software itself. TinkerCAD software assembled the premade parts quickly and with precision; the shape tool showed small details and sizes of the different objects for ease of assembly and several other tools made TinkerCAD a good software for novice assembly of the Open PCR machine.


Our Design

Screenshot 17).png

Screenshot 18).png


Aerial View

Screenshot (19).png


Close up of improvement

Screenshot 20).png


Our main improvement to the design is the removal of the knob that controls the cover of the PCR tubes. The knob and its shaft assembly was replaced with two latches, removing any doubt that the cap is on and securing the lid for the heating and cooling cycles. The original device was complicated in the intention of securing the PCR cap and the knobhead on the top of the thermal cycler had almost no usage and when turned uncapped several different components that were unrelated. The Open PCR machine designed here uses a much simpler method of latching to open the entire machine for ease of usage and control of the PCR operations.


Feature 1: Consumables Kit

The consumable kit that came with the original PCR reaction was fairly thoroughly prepared and well kept. The consumable kit that comes with this PCR machine will include the same products such as gloves, the micropipetters, SYBR Green solution, the master mix containing: taq polymerase, dNTPs, Buffer solution and primers; tray tables, micropipet tips and a disposable waste container. The consumables kit had several different problems with it during the experiment that added error and uncertainty to the experiment. The consumables kit that comes with our PCR machine includes an efficiently designed tray table that holds different test tubes of any size within the tray table itself. Furthermore, instead of wrapping the SYBR Green solution in an aluminium foil wrapping to keep it from deteriorating, the consumables kit comes with reaction tubes of SYBR Green that are all black on the outside in order to hold light sensitive solutions. Furthermore several reaction tubes will be included of different colors in order to differentiate between different samples, and on those tubes there will be rough strips to label the samples that are resistant against heat erosion. The tray table itself also has features of attaching to the actual PCR machine for ease of storage and space. The micropipettes have presets on the volume levels within this consumables kit. The micropipettor is now fully digitized so that the volume intake can be inputted electronically and distributed electronically with less human interaction at the press of a single putton. Preloaded pcr material is included in the forms of a master mix with all the materials combined and individual materials for further usage and combination of PCR products.

Feature 2: Hardware - PCR Machine & Fluorimeter

The PCR machine, while rather effective when the seal atop the machine is fully fastened, can show very skewed results if the hatch fails to be closed due to human error. With the current design, the latch may not fasten all the way because one must be sure to make sure the cap is all the way closed and the screw is completely tightened. Based on group observations of the machine, it is far too easy for this hatch to be improperly shut, thus causing the data to suffer accuracy. The proposed solution to such a problem is to diminish the likely hood of this human error. The group has decided to apply latches to the hatch atop the machine. These hatches will make sure that the top is both completely closed, as well as securely fastened throughout the duration of the PCR machine cycle.

As for the fluorimeter, the group found the greatest problem to be that the pictures taken of the droplets were not as accurate as they could be. With the way the front flap works, it is difficult for group members to place the camera inside the fluorimeter. Also, the flap fails to completely close, meaning that excess light is being let in which will affect imageJ results. The group proposed a possible curtain mechanism which will both make the fluorimeter easy to access and close better so as to keep light from entering.