Pattern-Recognition-in-Microbiomes

This repository aims to develop a web application to process and build a machine-learning model to geographically classify organisms at a taxonomical level.

Pre-requisites

Before running this app, make sure to install the required modules. Navigate to the requirements.txt file and run the code below in your suitable terminal environment

pip install -r requirements.txt

or

python -m pip install -r requirements.txt

After completing the above step, the user can download the other files and execute the command below to host the web app locally

python streamlit run app.py

Make sure the path is referenced properly.

Decription

The coded app contains four sections. Some sections are yet to be completed and hence warned about.

• Upload
• Profiling (incomplete)
• Modeling
• Download (incomplete)

Upload

This section is designed to process and model our dataset. Please ensure that you upload the OTUs, taxonomic, and metadata tables. Ensure that all columns in the sampled tables are named correctly. Use a single space (' ') as the separator for the OTUs and tax files, and use a tab space ('\t') for the metadata.

Choose the appropriate taxonomical hierarchy for classification for both the index and the target variable. This final processed dataset will be saved either on your local disk or on the server for subsequent sections. It can only be modified by pressing the merge button

Profiling

This section is planned to showcase a small EDA on the dataset the final processed dataset.

Modeling

In this section, the user is provided with the option to choose either

• To build a classification model of choice
• Compare all the classification models build around the default tuning parameters.

This is built with the idea that the user might want to identify the best modeling algorithm suitable for the dataset and later choose to examine that model's performance in the previous section.

Download

This section is designed to allow users to download the chosen model for the analysis (with default parameters), which can later be appropriately tuned by the user

Check out the webapp using this link Pattern Recognition