Kidney Dataset Visualisation

• 

  Main Page with 7 Sets of Images

• 

  SubPage with ROI of Set

• 

  ROI Image Details

• 

  PLot of Target Name and BioProbes

• 

  https://www.spandidos-publications.com/10.3892/mmr.2017.7666

• 

  https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4975268/

Inspiration

Many of my family members have Diabetes and hope this project can ease the researchers in the field so they can get a cure for Diabetes soon.

We know that it takes lot hardware resources (I just opened a single image from Dataset and my PC collapsed due to RAM overload) to visualise Slide Images in Biology due to its high resolution, so here I have made this WebApp which uses my optimised images from the NanoString Dataset and displays its Data easily and faster.

What it does

It is a WebApp developed using Python and Flask which optimises the dataset images and displays its features in form of graphs and plots with a easier and faster UI.

It shows major features of any ROI in Dataset like Normalization of Target Names, BioProbe Count and properties of the ROI.

How we built it

1. I first optimised and compressed the dataset images using Python so it will load faster on the Web (100MB to >1MB), and also converted the text dataset to csv files.
2. Then I created a Frontend in Flask with multiple pages to Display the Images in respective Order as Disease and Types.
3. Then I researched about major Genes for Diabetic Kidney Disease from Research Papers and filtered them out from 15000 Target Names using Pandas Library
4. Then similarly for BioProbes Value I mapped the Main Target Names filtered before and displayed using Graph Plot and the Dataset FIles "Sample Annotations: Kidney_Sample_Annotations.txt,Feature Annotations: Kidney_Feature_Annotations.txt,Probe Expression: Kidney_Raw_BioProbeCountMatrix.txt,Target Expression:Kidney_Raw_TargetCountMatrix.txt,Normalized Expression: Kidney_Q3Norm_TargetCountMatrix.txt".
5. Then Finally merged all components into Local Flask Server

Challenges we ran into

1. Plotting Bar Graphs was tricky as used the library for the first time
2. Using the Big Image files from dataset was hard due to less RAM in my computer

Accomplishments that we're proud of

1. Created Visualisation app with such a huge dataset for first time

What we learned

1. Image Handling in Dataset
2. Filtering Dataset's CSV files
3. Plotting Bar Graphs in Python

What's next for Kidney Dataset Visualisation

1. As the current UI is not so Good, so improving it using CSS and JQuery
2. Adding more Visualisations and Analysing them

Built With

• flask
• html
• plotly
• python