DeepSpot - DeepSpot is a novel deep-learning model that leverages recent foundation models in pathology to effectively predict spatial transcriptomics from H&E images. DeepSpot substantially improved gene correlations across multiple datasets from patients with metastatic melanoma, kidney, lung, or colon cancers as compared to previous state-of-the-art. Paper
aestetik - AESTETIK is a convolutional deep-learning model that jointly integrates spatial, transcriptomics, and morphology information to learn accurate spot representations. It significantly outperforms the current state-of-the-art on both structured and heterogeneous tissue samples across widely adopted technology platforms (e.g., 10x Genomics, NanoString Technologies, Inc.) enabling further progress in biomedical research applications. Paper
ezRun - An R meta-package for the analysis of Next Generation Sequencing data maintained and used by the Functional Genomics Center Zurich. Developed workflows for the analysis of single-cell data (e.g., integration of enrichR, gene enrichment, pathways activity, transcription factors binding, etc.).
Kipoi: Model zoo for genomics - This repository implements a python package and a command-line interface (CLI) to access and use models from Kipoi-compatible model zoo’s. Kipoi (pronounce: kípi; from the Greek κήποι: gardens) is an API and a repository of ready-to-use trained models for genomics. It currently contains 2194 different models, covering canonical predictive tasks in transcriptional and post-transcriptional gene regulation. Kipoi’s API is implemented as a python package (github.com/kipoi/kipoi) and it is also accessible from the command line or R.
kipoiseq - Standard set of data-loaders for training and making predictions for DNA sequence-based models.
gnomAD_DB - This package scales the huge gnomAD files (on average ~120G/chrom) to a SQLite database with a size of 34G for WGS v2.1.1 (261.942.336 variants) and 99G for WGS v3.1.2 (about 759.302.267 variants), and allows scientists to look for various variant annotations present in gnomAD (i.e. Allele Count, Depth, Minor Allele Frequency, etc. - here you can find all selected features given the genome version). (A query containing 300.000 variants takes ~40s.)
hail-db - This package allows the user to inject VCF files in Hail-Tables and query VCF-tables for GT, GQ, DP fields for variants and samples of interest. The advantage is that you can scale huge VCF files and query them efficiently on a machine with limited computational resources.
Innovation Challenge 2020 – Siemens, 2020 Developed a gamification application with the aim of reduction of the daily automotive carbon footprint
hackaTUM C0dev1d19 - During the Corona (COVID-19) pandemic it became obvious that a lack of healthcare workers and equipment can cause a severe impact in the fight against the virus. The rapidly growing shortage of ventilators is becoming of dire concern. Our platform provides intuitive interface which shows real time ventilator information for multiple patients on the same place, so that the medical workers can monitor the health status of their patients easily on the go.