A single droplet of blood contains more information than any human could ever sort through by hand — tens of thousands of genes, each switching on or off in millions of individual cells. That much data would fill 10,000 dimensions, a space so vast no one can picture it. But a team at the University of Basel in Switzerland has built a tool that finally makes sense of it, and it has already revealed a type of immune cell that science didn't know existed.
The problem starts with how modern biology works. Over the past decade, science entered the age of "big data." Using a technique called single-cell RNA sequencing, researchers can measure the activity of tens of thousands of genes inside hundreds of thousands — even millions — of individual cells at once. This data promises to reveal how cells develop, communicate, and change during disease. The catch? Interpreting it.
"People are good at recognizing patterns in two or three dimensions," says professor Erik van Nimwegen. "But we simply can't make a picture of a dataset that exists in 10,000 dimensions."
Most existing tools try to flatten that towering mountain of data into a flat, two-dimensional map. These maps are used in nearly every study, but they distort the data — making it impossible to tell whether the relationships shown between cells are real or just an illusion created by squeezing 10,000 dimensions into two.
The Basel team's software, called Bonsai, takes a different approach. Instead of a flat map, it builds a branching tree, with individual cells sitting at the tips of the branches. "Crucially, the distances along the branches accurately reflect how closely cells are related in the high-dimensional space," says first author Dr. Daan de Groot.
The team tested Bonsai on both simulated and real datasets. Compared with existing methods, it reconstructed developmental pathways far more accurately, preserved true relationships between cells, and identified similar cells much more reliably. The results were published in the journal Nature Biotechnology.
Then came the discovery. When the researchers analyzed human blood cells, Bonsai not only automatically recovered the known relationships between different blood types — it also uncovered a previously unknown subtype of a kind of immune cell called a natural killer (NK) cell. The molecular signature of this new subtype showed it comes from a lineage called the myeloid lineage, completely different from the lymphoid lineage that all NK cells were thought to come from.
"This example shows why faithful representation and visualization of complex data matter," says van Nimwegen. "When you can trust the picture, you have a much better chance of making new discoveries."
Because Bonsai can handle any kind of high-dimensional data — not just gene expression, but medical records, the mix of species in a microbe community, or the firing patterns of neurons — the researchers see it as a tool for scientists across all of biology. And it's freely available to researchers everywhere. The next hidden pattern in the data may be one discovery away.