remylau/embiggen

🍇 Embiggen is the Python Graph Representation learning, Prediction and Evaluation submodule of the GRAPE library.

Hi there 👋

I am a Postdoctoral Fellow in the Ray and Stephanie Lane Computational Biology Department at Carnegie Mellon University. I currently work on developing computational and machine learning methods, as well as software, for analyzing and understanding single-cell epigenomics.

I obtained my Ph.D. degree in the Department of Computational Mathematics, Science & Engineering (CMSE) at Michigan State University. My training focused on network biology, graph representation learning, spectral graph theory, and machine learning.

🛠 I’m currently developing

• Stay tuned! Something exciting is going online soon!

🧰 I'm actively maintaining several packages related to my past / recent projects

• obnb: a Python toolkit for setting up benchmarking datasets using publicly available biomedical networks and gene annotation resources. A comprehensive benchmarking study with various graph neural networks and graph embedding methods is presented in obnbench.
• DANCE: an extensive toolkit for deep learning with single-cell (multi-)omics data.
• PecanPy [paper]: a memory efficient and Numba accelerated Python implementation of node2vec with an improved version node2vec+ [paper] for weighted graphs.
• PyGenePlexus [paper]: a network-based gene classification service using machine learning and gene interaction network features.
• GTaxoGym [paper]: a taxonomic study of benchmarking graph datasets from various domains based on the GNN model sensitivity to a collection of graph perturbations.

📫 Find out more about my work and reach out to me

• ORCID
• Google Scholar
• Twitter
• Email: [email protected]

⚡ Side projects

• ✍️ I share my passion about network biology and machine learning via blog posts on Medium
• 👀 I create mathematical and algorithmic visualizations using Manim, which was first developed and used by my favorite math YouTube channel 3Blue1Brown.
• 🤗 I contribute to open source projects in various ways
  • Implemented patches and fixes: Pytorch Geometric
  • Feature discussion: Hydra
  • Bug report and question posting: mypy, mydisease.info, ndex2-client
• 📦 I work on several small packages on the side to help improve my production workflow and exercise my dev workflow
  • pydab: a tool for working with dab files used by Sleipnir, a C++ library for machine learning on genomic data.
  • py2zenodo: a command line tool for uploading data to Zenodo
  • fastauroc: a Numba accelerated computation of the area under the receiver operating characteristic.

Installation notes

conda create -n remylau python=3.11 -y && conda activate remylau

pip install -e .

conda clean --all -y