I am a Senior Data Scientist at BRIGHT, the Novo Nordisk Foundation Biotechnology Reasearch Institute for the Green Transition (formally DTU Biosustain - the Novo Nordisk Foundation Center for Biosustainability) in Copenhagen.
You can find nearly all of my work on GitHub under the organisations DTU Biosustain, the Multi-Omics-Analytics-Group (MONA), and RasmussenLab.
Currently, I'm working on metabolomics and (phospho-) proteomics datasets in the context of engineering cell factories. I also worked on self-supervised deep learning models for MS-based proteomics imputation. See the PIMMS repository and the Nature Communications paper. You might also be interested in checking out a comparison based on PIMMS on an Alzheimer's dataset: rasmussenlab.github.io/pimms/
For an easy comparison of proteomics data and some clinical metadata, have a look at not just another biomarker (njab). You can easily run the example notebook on Colab (which is regularly and automatically tested to work) and plug in your data. Having the associated Python package will make it easy to extend the experiments. The code was used in a Nature Medicine paper and a Scientific Reports paper.
The current effors are centered around an ecosytem of acore, vuecore and vuegen:
- acore is Python package to preprocess and analyse multi-omics data.
- vuecore is Python package for creating interactive and static visualizations of multi-omics data, that builds default plots on top of outputs from acore which can be augmented by users.
- VueGen automates the creation of reports from bioinformatics outputs, supporting formats like PDF, HTML, DOCX, ODT, PPTX, Reveal.js, Jupyter notebooks, and Streamlit web applications. Users simply provide a directory with output files and VueGen compiles them into a structured report.
Additionally we maintain a package for growthcurves in Python:
- growthcurves calculate with different methods (models, sliding window, etc.) the most relevant growth curve parameters
which is then used in an evolving collection of apps for specific scientific setups. See an overview:
- common Data Science Platform notes
- An easy-to-use notes template
- Lab website for technical notes (based on Sphinx, deployed via GitHub)
- ProteoBench - focus on software parameter file parsing
ProteoBench is an open and collaborative platform for community-curated benchmarks for proteomics data analysis pipelines. Our goal is to allow a continuous, easy, and controlled comparison of proteomics data analysis workflows.
- Creating and curating a Python package template