Lücken Lab

At the Lücken Lab, we build machine learning models on single-cell data and translate these to clinical applications predominantly in lung research. In particular, we focus on building cellular reference models of human tissues by integrating diverse single-cell datasets to improve the representation of human diversity. Using these reference models we are pioneering efforts to model patient variation, enable rapid analysis of new data, and work towards personlized medicine

Follow us on X

At the Lücken Lab, we build machine learning models on single-cell data and translate these to clinical applications predominantly in lung research. In particular, we focus on building cellular reference models of human tissues by integrating diverse single-cell datasets to improve the representation of human diversity in health and disease. Using these reference models we are pioneering efforts to model patient variation, enable rapid analysis of new data, and work towards applications in personlized medicine.

Follow us on X

Our Research

Our research so far has focused on building atlases, using these atlases, and designing quality standards for single cell analyses.

Our mission is to translate single-cell tools and machine learning methods to clinical applications, with a specific focus on atlases and pulmonary research. High standards in terms of robustness and reliability are set for methods used in clinical practice. To promote translatability of our methods, we aim to apply these standards throughout our single-cell work by promoting open, community-driven benchmarking:

Determine single-cell best practices
Better understand patient variability in pulmonary diseases
Derive relevant clinical/translational insights from single-cell data

Our Values

Scientific Values:

Dare to innovate: When beneficial, try a new approach instead of the tried and true that is limited. Leverage the lab’s expertise to test new approaches that solve that challenges you are facing
Science before ego: We want to produce science that is robust, reproducible and open. If the data shows we are wrong, that is okay. We take time to ensure usability of our tools so that scientific goals can be achieved also outside our lab.
Constructive criticism is caring: We are our own strongest critics. We aim to ask for feedback from our peers, and we are willing to be forthcoming with our own feedback. We value the time spent by others to help us improve our work.

Interpersonal values:

Communicate with respect: Communication is central to our lab culture as it is the instrument how we make each other better. To ensure the information we want to convey is received well, we strive to choose our words in such a way that treats our peers with respect.
Diversity in thinking and representation: Scientific challenges are complex. This complexity is best tackled from multiple points of view. Therefore, bringing together diverse ways of thinking is crucial for scientific progress.
Have fun: If science isn’t fun, we’re doing something wrong!

Business, Technology, Internet and network concept. Core values responsibility ethics goals company concept. 3d illustration

SPatial Analysis of Cancer Evolution in the Tumour Immune MicroEnvironment (SPACETIME) is an EU consortium studying how spatial cellular organization changes with disease severity in non-small cell lung carcinoma (NSCLC). We are collaborating with biologists, clinicians, biotechnologists, and other computational biologists or machine learning scientists to build a spatial multiomic map of NSCLC across species and disease stages with the goal of building better diagnostics, expanding our understanding of disease mechanisms, and developing treatment options to improve outcomes for lung cancer patients.

We are heavily involved in the Human Cell Atlas (HCA) and are in a leading role in the HCA integration team. Here we coordinate an international group of PhD students and postdocs to build the first version of the HCA, a cellular reference atlas for each human tissue or organ. The integrated Human Lung Cell Atlas (HLCA) represents the first large-scale, integrated single-cell reference atlas of the human lung. It consists of over 2 million cells from the respiratory tract of 486 individuals and includes 49 different datasets. The HLCA is widely used in the community, with over 2,000 downloads of the model alone.

Respiratory Precision Medicine PhD Training Network- Excellence through Systems Biology, Spatial and Single-cell Transcriptomics.
We are participating in an EU-funded doctoral network on understanding mechanisms, identifying biomarkers, and developing treatments for chronic lung diseases including Asthma and COPD. We interact with groups across disciplines to build tools and analyse data to generate insights in asthma and COPD with translational impact.

Benchmarking formalized challenges in single-cell analysis:
Our goal is to provide an open source, community driven, extensible platform for continuously updated benchmarking of formalized tasks in single-cell analysis.

Single-cell datasets are growing in size, and slowly also in the number of samples profiled. Additionally, integrated reference atlases that contain hundreds to thousands of samples are becoming increasingly available. This enables studying inter-individual variation from single-cell datasets. We have pioneered the study of inter-individual variation on single-cell data by aggregating datasets, extended this to understand the effects of demographic variation on human lung transcriptomes, and are developing and benchmarking approaches that represent inter-sample variation.