BME103 s2013:T900 Group4 L3

From OpenWetWare
Revision as of 07:32, 16 April 2013 by Amelia A. R. Lax (talk | contribs) (New System: Design Strategy)
Jump to: navigation, search
Owwnotebook icon.png BME 103 Spring 2013 Home
Lab Write-Up 1
Lab Write-Up 2
Lab Write-Up 3
Course Logistics For Instructors
Wiki Editing Help
BME494 Asu logo.png


Name: Kinjal Ahir Role:Protocol
Name: Zach Young
Initial Machine Testing
Name: Anna Essex
Initial Machine Testing
Name: Tuan Phan
Research and Design
Name: Amelia Lax
Research and Design


Original System: PCR Results

PCR Test Results

Sample Name Ave. INTDEN* Calculated (μg/mL) Conclusion (pos/neg)
Positive Control 1,450,385 1.60 pos
Negative Control 488,789 0.18 neg
Tube Label: B2 Patient ID: 17818 rep 1 709,603 0.51 pos
Tube Label: C2 Patient ID: 17818 rep 2 642,405 0.41 pos
Tube Label: D2 Patient ID: 17818 rep 3 417,721 0.07 neg
Tube Label: B1 Patient ID: 85158 rep 1 450,174 0.12 neg
Tube Label: C1 Patient ID: 85158 rep 2 387,850 0.03 neg
Tube Label: D1 Patient ID: 85158 rep 3 376,360 0.01 neg

* Ave. INTDEN = Average of ImageJ integrated density values from three Fluorimeter images

Bayesian Statistics
These following conditional statistics are based upon all of the DNA detection system results that were obtained in the PCR lab for 20 hypothetical patients who were diagnosed as either having cancer or not having cancer.

Bayes Theorem equation: P(A|B) = P(B|A) * P(A) / P(B)

Calculation 1: The probability that the sample actually has the cancer DNA sequence, given a positive diagnostic signal.

  • A = frequency of cancer-positive conclusions = 9 / 20 = 0.45
  • B = frequency of positive PCR reactions = 26 / 60 = 0.43
  • P (B|A) = frequency of positive PCR given cancer-positive conclusion = 24 / 26 = 0.92
  • P(A|B) = 0.96 = 96%

Calculation 2: The probability that the sample actually has a non-cancer DNA sequence, given a negative diagnostic signal.

  • A = frequency of cancer-negative conclusions = 11 / 20 = 0.55
  • B = frequency of negative PCR reactions = 34 / 60 = 0.57
  • P (B|A) = frequency of negative PCR given cancer-negative conclusion = 31 / 34 = 0.91
  • P(A|B) = 0.88 = 88%

Calculation 3: The probability that the patient will develop cancer, given a cancer DNA sequence.

  • A = frequency of "yes" cancer diagnosis = 9 / 20 = 0.45
  • B = frequency of "pos" test conclusion = 26 / 60 = 0.43
  • P (B|A) = frequency of pos given yes = 24 / 26 = 0.92
  • P(A|B) = 0.96 = 96%

Calculation 4: The probability that the patient will not develop cancer, given a non-cancer DNA sequence.

  • A = frequency of "no" cancer diagnosis = 11 / 20 = 0.55
  • B = frequency of "neg" test conclusion = 34 / 60 = 0.57
  • P (B|A) = frequency of neg given no = 31 / 34 = 0.91
  • P(A|B) = 0.88 = 88%

New System: Design Strategy

We concluded that a good system Must Have:

- Easily Determined Results: The easier the results are to read accurately, the less likely a misdiagnosis in either direction. It is undesirable both to give a false negative, where a patient is not treated when care is needed, or to give a false positive, wasting time and resources on those who do not need them. This aspect is central to any diagnostic tool.

- Simple OpenPCR Software: Simplicity increases ease and efficiency in lab experiments and hopefully leads to faster diagnoses. It also makes troubleshooting easier should problems arise. The more straightforward the system, the more quickly users can learn to use the machine.

We concluded that we would Want a good system to have:

- Low Cost: Currently an OpenPCR machine costs $599 and a fluorimeter costs $300. An inexpensive material would help reduce cost and increase accessibility, since there is always a limited budget for new equipment. This would not only allow users to increase the amount of tests that can be run at the same time, but also boost sales, which is important for marketing any device.

- Integrated Camera: Phone cameras are easily moveable and vary in size and quality, leading to differing results. Adjusting smartphone camera settings can be time consuming or, in some models, impossible. Having a built-in camera increases cost, but it is worth it to increase speed and accuracy. This also simplifies the software is because the PCR does not have to adjust to different cameras. Finally, phone sizes and shapes vary enough to make building a universal cradle to fit them difficult.

We concluded that a good system Must Not Have:

- Troublesome USB Connectivity: USB connectivity should function well in order for the PCR machine to work. This also reduces troubleshooting time and is a fairly simple problem to fix in a new system's design.

- Flammable Casing: The PCR rapidly cycles through different temperatures, some extremely high. As with other electronics, the wires and other electrical components could overheat or spark. At the very least, the wooden casing currently used could damage the machine beyond repair, and at worst, lead to massive fires in areas with expensive equipment and many people. Simply changing the casing to a material like plexiglass would solve this problem.

We concluded that a good system Should Avoid:

- Slow Amplification: Increasing the speed of diagnosis would make the test more efficient and possibly allow for a verdict while the patient is still in the doctor's office without spending an inordinate amount of time. However, speed will be sacrificed if need be to accommodate more important features such as accuracy.

- Movable Phone or Fluorimeter: In the current system, which uses a smartphone camera in a cradle, the phone or fluorimeter can be easily moved by accident, which requires readjustment and can possibly skew results. Using the current setup decreases accuracy but also decreases cost. We placed accuracy as more important than cost, but the opportunity to reduce price if the machine is much too costly keeps this idea from being completely rejected.

New System: Machine/ Device Engineering


Current design of fluorimeter

Rather than drastically change a fairly-efficient PCR machine, we decided that the fluorimeter setup was more in need of modification. The only change to the PCR machine would be improved USB ports, but the fluorimeter would have a built-in camera to remove the complications of positioning a camera phone. The phone would still be used to run the machine, but it wouldn't directly take the pictures. This new camera would take the place of the current cradle and be at a fixed position in respects to the fluorimeter for most efficient photographing. Also, the slots on the board of the fluorimeter would be labeled to avoid confusion in the process of analysis.


Fluorimeter - We chose to include these new features:

  • Integrated Camera - helps reduce inconsistency of photography and time-consuming difficulty of positioning
  • Labeled slots - reduces likelihood of error from misidentified photographs

PCR Machine - We chose keep these features the same as the original system:

  • Reliable Hardware - the machine is sturdy and does its job efficiently considering its simple construction
  • Preexisting Software - the current Open PCR software is well developed and user-friendly


  • Step 1: Connect the camera unit to the fluorimeter.
  • Step 2: Adjust the camera settings according to the current experiment.
  • Step 3: Link the camera to the phone being used to control the experiment.
  • Step 4: Take photo.
  • Step 5: Upload photo for necessary manipulation.

New System: Protocols

We chose to modify the hardware of the fluorimeter. However, overall protocols should remain the same. As the PCR machine was not modified, its protocols will also remain unaltered.


Supplied in the Kit Amount
Camera Unit 1
Reaction mix 400μL
Battery 1
Software freeware
Supplied by the User Amount
Filter water 1,000μL
SYBR Green 2,000μL
Primers 4,000μL
DNA sample (negative and positive) 400μL


  • PCR Protocol
  1. Obtain and label 8 50μL DNA samples (4 each from 2 patients, positive control, negative control) and 8 50μL tubes of PCR reaction mix
  2. Set micropipette to 75μL and attach disposable tip
  3. Transfer all of the liquid from positive control DNA sample to a reaction mix tube, discard tip, label tube
  4. Repeat for the remaining 7 DNA samples
  5. Set the PCR program to run three stages
Stage Number of Cycles Temperature (°C) Duration
1 1 95 3 minutes
2 35 95 30 seconds
n/a 57 30 seconds
n/a 72 30 seconds
3 n/a 72 3 minutes
Final hold n/a 4 n/a
6. Load the mixed tubes into the PCR machine, close the lid, run the program
7. Remove tubes at end of program

  • DNA Measurement and Analysis Protocol
  1. Obtain a tray of sample tubes (8 buffer, 2 SYBR GREEN, 1 H2O, 5 calf Thymus DNA, 8 PCR reaction samples)
  2. Set micropipette to 120μL and attach disposable tip
  3. Transfer all of the liquid from positive control PCR sample to a buffer tube, discard tip, label tube
  4. Repeat for the remaining 7 PCR samples
  5. Calibrate the fluorimeter using the calf thymus DNA samples of known concentration
    • Set up the integrated camera and adjust the settings, connect the smartphone to the camera for control
    • Place 80μL of SYBR GREEN I onto the slide so it forms a definite drop, add 80μL of DNA, align with the LED
    • Cover the fluorimeter with the light box and use phone to take picture
    • Remove the box, remove the drop, export picture to computer for analysis (Make sure to label!)
  6. In ImageJ, adjust the settings and split the color channels of the image to select green
  7. Draw an oval around the image and measure, repeat for background
  8. Repeat for the remaining concentrations of calf thymus DNA
  9. Use these readings and the known DNA concentrations to create a graph with a linear fit for calculating concentration based on INTDEN values
  10. Using the same procedure, repeat with the unknown DNA samples from the patients and use the graph to calculate DNA concentration

New System: Research and Development


CHEK2 is a gene located at chromosome 22. It provides instructions for making protein called checkpoint kinase 2, a tumor suppressor. This particular protein responds to damage in DNA, preventing the cell from entering mitosis when the cell's DNA deviates from normal. Mutations of CHEK2 gene can lead to breast cancer, Li-Fraumeni syndrome, and other type cancers and diseases.

In our design, we chose to use primers for both normal and cancer-associated DNA sequences so users can cross-check their results. For example, if a patient's test returns positive for the cancer-associated allele, the test can be run with a normal allele primer to ensure that the test results were accurate, in which case the test for the normal allele should be negative. This addition of primers associated with the normal sequence allows users to test for the presence of normal DNA if they suspect that the cancer-associated DNA is not present. This helps fulfill our goal of making a more reliable test. It takes more time to run more samples, but it provides an extra layer of caution to help ensure correct diagnosis.

Primers for PCR
Normal Allele

Cancerous Allele

Our primers address the following design needs

  • accuracy of diagnosis: As we considered diagnostic accuracy absolutely essential to a new system,
  • Design specification 2 - explanation of how an aspect of the primers addresses any of the specifications in the "New System: Design Strategy" section
  • Etc.

New System: Software

As has been seen by the several groups who already have software in development, the need for more efficient PCR and image analysis capabilities are growing. For our particular machine, an app allowing a smartphone to control the integrated fluorimeter camera would be most essential, and ideally this app could also perform image analysis, lessening the complication of transferring large quantities of images that all look very similar to the human eye.

As ImageJ is a Java-based program, converting it to an Android app would be fairly straightforward. Similarly, channeling a phone to control an external camera is nothing particularly revolutionary but would greatly increase the efficiency and accuracy of this lab. Open PCR software could also be streamlined to run more quickly and give more accurate readouts during runtime. In reality, software modifications would provide the most improvement for this lab with relatively little effort.