Project IE is a python supported module which parses, stores and analyses cochlear MRI data from text files. This prototype currently allows the user to input one raw file at a time to be tidied and passed through pandas DataFrame for easy data management. The resulting DataFrames also enable further processing in the form of a one-way ANOVA analysis or can alternatively output a graph which is then saved as an image (.png).
Visualisation (graphing) was achieved using a plotting library called matplotlib and the one-way ANOVA analysis was completed using the SciPy library
This study looks to quantify spatial and temporal inflammation-induced changes in the capsular permeability and macrophage infiltration in guinea-pig cochlea using MRI. Modelling of such exchanges in blood and different inner ear (IE) compartments require the analysis of a substantial amount of data.
This data was obtained from an ongoing replication study currently being conducted at the University of Auckland. Using Project IE, the MRI values can be extracted and indexed. Each text file contains a set of MRI which measures the propagation of a contrast agent injected into the IE.
Although a prototype, this project could possibly provide reference data that can in future be used to quantitatively assess the treatment of ear disease in animal models and establish a platform from which such techniques can be transferred into clinical practice.
##Contributors
- Bibiana Lee
- Kreshnik Pireva
- Callum Chalmers
-
Python
- Successful running of this programme will require Python 2.7.12
- You can download Python 2.7.12 following This link
-
Operating System
- Ubuntu 16.04 LTS
- Python Packages
- Python 2.7.12
- Instructions on how to install can be found [here] (https://wiki.python.org/moin/BeginnersGuide/Download)
- pandas version : u'0.18.1
- Instructions on how to install can be found here
- matplotlib 1.5.3
- Instructions on how to install can be found here
- SciPy 0.18.1
- Instructions on how to install can be found here
- Python 2.7.12
##Data File Requirements
- Data Files must be text files (.txt)
- Files should be labeled like this: (no.of images)D(no.days after treatment or non-treatment)
- So the name should contain a number followed by a D then another number
- For example: a file labeled 13D4 is an animal that has been imaged 13 times, 4 days after treatment
###Labels - THE SCRIPT WILL NOT WORK WITHOUT THESE LABELS
-
The data file must contain the following labels. This is a must.
-
The Ktrans and Diffusion coefficient labels must be followed by four numbers
-
ID1 can be either 294 (Right ear) or 295 (Left ear)
-
ID2 is the height of the cochlear anatomy and ranges from 0 - 7
## INPUTS ## # arterial input function values aif: # time after fixed capture, in second time: ## OUTPUTS ## **k trans value, average value, std value, max and minimal values** local-cochlea-bm.ktrans_ID1_ID2:val_mean,val_std,val_max,val_min **Pixel enhancement average value along the time** local-cochlea-bm.enhan(t)_ID1_ID2:enhancement1,... **Pixel enhancement std value along the time** local-cochlea-bm.std_enh(t)_ID1_ID2:std_enhancement1,... **Diffusion coefficient, average value, std value, maximal and minimal values** local-cochlea-bm.difcoef_ID1_ID2:coef_mean,coef_std,coef_max,coef_minExample Raw Data
aif,0,0.914,1.115,1.235,1.225,1.218,1.122,1.106,1.192,1.017,1.214time,0,120,220,320,420,520,620,820,920,1120,1220local-cochlea-bm.mha_ktrans_294_0,0.0123,0.0025,0.0326,0.00425local-cochlea-bm.mha_enhanc(t)_294_0,3.926,5.013,4.329,4.426,5.036,4.627,4.858,3.364,4.764,4.247local-cochlea-bm.mha_std_enh(t)_294_0,2.015,2.706,2.263,2.251,2.696,2.342,2.487,1.985,2.219,2.405local-cochlea-bm.mha_difcoef_294_0,0.0123,0.0025,0.0326,0.00425
##Running the programme using the test data ###Loading project_IE.py
-
Download the
DummyData_Project_IEfolder and transfer the project_IE.py into this folder.- This step is incredibly important as the file will not run without the datafiles present in the folder
-
Open a command line terminal in the folder containing project_IE.py
-
Type in python project_IE.py in the command window
-
Once this is complete you should be greeted with the first menu
Welcome to Project IE your options are: 1) One file 2) Quit Choose your option:If you would like to exit the programme you can type
2otherwise keep reading the instructions
###Loading dummy data
- Type
1
-
If you enter an invalid number a message will appear saying
The number you have entered is invalid please choose either option 1 or 2
-
Type in
10D4- Your DataFrames should print now this should be followed by a message that reads:
Your graph has been plotted and saved You may choose another option now -
The secondary menu should greet you now
What would you like to do now? 1) visualisation 2) one-way ANOVA 3) Quit Choose your option:
###Visualisation
- Type
1for a visualization - A graph should automatically save in your folder called
Your Graph - A new message will appear
Your graph has been plotted and saved. You may choose another option now. - Once this is complete you will be greeted by the secondary menu again
-
Type
2for a one-way ANOVA -
A message should appear saying:
One-way ANOVA followed by your P value -
An interpretation of your p-value should this output
- If it is significant (p-value <0.05)
The differences between some of the means are statistically significant Your analysis is complete. You may choose another option now.- If it is not significant (p-value >0.05)
The differences between the means are not statistically significant Your analysis is complete. You may choose another option now.Once complete you can now exit the programme by typing
3
-
The expected output for the dummy data should look like this for the Ktrans DataFrame
-
A seperate text filed called
10D4_expected_output.txtcan be found in the DummyData_Project_IE folder to compare your results--------------------------------------------------- Your Ktrans DataFrame --------------------------------------------------- Mean STD Max Min Ktrans ID1 ID2 ktrans 294 0 0.0123 0.00250 0.0326 0.00425 1 0.0115 0.00450 0.0224 0.00354 2 0.0121 0.00740 0.0422 0.00228 3 0.0098 0.00470 0.0253 0.00177 4 0.0187 0.00520 0.0309 0.00326 5 0.0121 0.00290 0.0468 0.00292 6 0.0073 0.00320 0.0103 0.00865 7 0.0041 0.00250 0.0276 0.00342 295 0 0.0123 0.00125 0.0326 0.00425 1 0.0118 0.00664 0.0133 0.00242 2 0.5411 0.00896 0.0505 0.00601 3 0.8471 0.00843 0.0554 0.00808 4 0.0676 0.01116 0.0939 0.00162 5 0.0199 0.00763 0.0045 0.00496 6 0.0131 0.00923 0.0889 0.00411 7 0.0024 0.00616 0.0385 0.00315
The expected graph output can be found in the DummyData_Project_IE folder and is in the Expected_Outputs folder in a .png format.
###one-way ANOVA
-
The expected output for the dummy data should look like this:
___________________________________________________ One-way ANOVA P = 0.562677075784 The differences between the means are not statistically significant Your analysis is complete. ___________________________________________________ You may choose another option now.
##Limitations
- The analysis and graph produced are for demonstrative purposes only. This is a prototype.
- This prototype only checks for
#andempty spacesas the MRI data text files not contain any other intrusive characters. Other characters such as a@may not detected by the software while parsing. It is therefore advised that the user manually checks the file first if they are using their own data. - The graph output is labeled "Your Graph" and will need to be manually re-named
- Once a file is in the programme it will be parsed and analyzed and it must run its course before you can quit and load another separate file
- This prototype is specifically for MRI text datafiles produced by an ongoing study so the labels must be present in the raw data or else the code will not work
- This programme has not been tested in other environments
- Programme testing requires a
pytest version 2.9.2which can be downloaded and installed following this [link] (http://doc.pytest.org/en/latest/getting-started.html) - As this programme has not been tested in other environments the developer cannot guarantee its use in Windows or OS.
- This test has only been run in a Linux environment
- In order to test the script please ensure that the
Project_IE.pyis in theTesting_Project_IE folder - Open
Project_IE.pyand add a#in front ofmenu()on the last line - Go to line 74 and remove the
#in front ofreturn lines - Open command in the folder
Testing_Project_IE folder- a comphrehensive guide on command line can be found [here] (https://help.ubuntu.com/community/UsingTheTerminal) - Once here type in
-m pytest - You should see the following results
============================= test session starts ============================== platform linux2 -- Python 2.7.12, pytest-2.9.2, py-1.4.31, pluggy-0.3.1 rootdir: /home/admin736/Desktop/BibianaLee-medsci736/Testing_Project_IE, inifile: collected 9 items test_project_IE.py ......... =========================== 9 passed in 0.49 seconds =========================== admin736@UOA346266:~/Desktop/BibianaLee-medsci736/Testing_Project_IE$ - 9 tests should have passed
- If these tests do not pass please make sure that the
menu()has a#in front of it and the#in front ofreturn lineshas been removed - If the tests pass please removed the
#in front of themenu()and reinstate the#in front of thereturn lines
- **Now you can run the programme using the instructions above using the Dummy Data provided or with you own file ** Please Note
- If you are using your own data be aware of the data requirements or feel free to modify the code to suit your needs
- Simply follow the instructions outlined above.
- Given that the tests run and you have completed a test run with the dummy data you should be able to navigate your way around Project IE
- Errors will be displayed and if there are any problems feel free to contact me through GitHub’s issue system
##Acknowledgement
- Code was taken and modified from Callum Chalmer (2016) and was permissible under the MIT licence
- Pathway codes and some test code was used as a foundation for Project IE
- Callum Chalmer (2016)
- Title: ePygenetics
- Author: Callum Chalmer
- Date: 2016
- Code version 1
- Availability: https://github.com/UOA-MEDSCI-736/CallumChalmers29-crispy-disco
```