Bonsai: Exploratory Analysis of Single-cell Data

Attendance is free of charge; however, registration is mandatory.

Overview

Single-cell omics data hold great promise for understanding complex biological processes, but exploring these data is complicated by their high-dimensionality and the dominance of measurement noise. Standard exploration methods, like UMAP and t-SNE, embed the data into two dimensions, which provably distorts structures in the high-dimensional data and even hallucinates new structures. The consensus in the field seems to be that these tools are merely used because there is no better alternative.

An alternative, Bonsai (preprint), reconstructs the most likely tree structure that relates high-dimensional objects while rigorously accounting for heterogeneous measurement noise. Unlike other visualization tools, Bonsai preserves true distances between objects and automatically regularizes noise. We also introduce the accompanying online tool, Bonsai-scout, which enables interactive exploration of the tree structure.

In this 0.5-day course, we will combine an explanation of the theoretical underpinnings of Bonsai, with hands-on exercises in which we show how to reconstruct a Bonsai-tree on your own data, how to interpret the results in Bonsai-scout and, finally, how to turn these into publication-ready figures.

Audience

This course is designed for PhD students, postdoctoral and other researchers in the life sciences from both academia and industry who are interested in analysing high-dimensional single-cell data.

Learning outcomes

At the end of the course, participants are expected to:

Explain the Bonsai method and its advantages for high-dimensional data exploration.
Apply Bonsai to construct trees from high-dimensional data.
Explore and interpret Bonsai results using the Bonsai-Scout web server.
Use and adapt downstream analyses of examples in Jupyter notebooks to generate publication-ready figures from Bonsai output.

Prerequisites

Knowledge / competencies

This course is designed for beginner level users. Participants should be familiar with single-cell data.

During the course, we offer participants the opportunity to ask questions about the analysis of their own data. In the case of scRNA-seq data, it is recommended to upload your scRNA-seq data at least a week in advance to the Bonsai-Scout web server at https://bonsai.unibas.ch.

To facilitate the hands-on experience with tree reconstructions, participants should already clone the Bonsai GitHub repository at https://github.com/dhdegroot/Bonsai-data-representation and follow the instructions in the ReadMe to install the right Conda environments. We recommend that participants meet the learning outcomes of Version Control with Git.

Technical

Participants are required to have their own computer with an Internet connection.

Application

Attendance is free of charge; however, registration is mandatory. Applications will close on 18/03/2026.

You will be informed by email of your registration confirmation.

Please note that participation in SIB courses is subject to our general conditions.

Venue and Time

This course will be streamed using Zoom.

It will start at 14:00 CET (08:00 EST) and end around 18:00 CET (14:00 EST).

Precise information will be provided to registered participants in due course.

Additional information

Coordination: Monique Zahn, SIB Training group.

You are welcome to register to the SIB courses mailing list to be informed of all future courses and workshops, as well as all important deadlines using the form here.