Welcome to CSE 527 (Computational Biology)! To help everyone brainstorm final project ideas, we've provided some sample datasets that have previously been used in high impact computational biology publications.
The repository is organized as follows:
In this folder, each top-level folder represents a data modality (e.g. gene expression data), with each subfolder corresponding to a specific dataset. Each dataset folder contains a single Jupyter notebook. Jupyter is an interactive computing environment that allows you to run Python code in your browser and interact with the output of or code. See here for instructions on how to install Jupyter and here for instructions on how to run Jupyter after installation.
For each dataset, the corresponding notebook provides a high level description of the dataset, and then walks through the process of downloading the dataset and performing any necessary preprocessing. Ideally, the notebooks are meant to encapsulate the whole process of obtaining/cleaning the dataset (i.e., the notebook should download any files rather than forcing the use to do it externally). However, due to restrictions on data sharing for some datasets, you may need to download the files yourself from another source.
TDC is an open-science platform with AI/ML-ready datasets and learning tasks for therapeutics, spanning the discovery and development of safe and effective medicines. TDC also provides an ecosystem of tools, libraries, leaderboards, and community resources, including data functions, strategies for systematic model evaluation, meaningful data splits, data processors, and molecule generation oracles. All resources are integrated and accessible via an open Python library.
This folder contains example script of how to use MIMIC-III, a freely-available electornic health record database comprising deidentified health-related data associated with over 40,000 patients who stayed in critical care units of the Beth Israel Deaconess Medical Center between 2001 and 2012. It will take few steps to obtain data access here which will only take few days.
eICU is another publicly available EHR dataset. Same as MIMIC, uses would need to requesting access to the eICU Collaborative Research Database through here
This folder contains example code for Chest x-ray images [https://physionet.org/content/mimic-cxr/2.0.0/]
Please reach out to the TAs with any questions! The initial datasets in the repository are datasets that the TAs are familiar with from previous research projects, and we'll be adding more over the next few days. If you're looking for a specific kind of data but can't find it here, please get in touch; we're more than happy to point you in the right direction.
If you already have data (e.g. from your own independent research) and are comfortable sharing it, you can earn participation points for the class by contributing your dataset to the repository. Please get in touch with the TAs for instructions on how to contribute.
Please go to office hour or contact TAs if you have any question.