You are here

Finding the odd cell out

Researchers from IIT and IIIT Delhi design an algorithm to find rare cells 

Our body is made up of trillions of cells—comprising of all shapes, sizes and kinds. Among this vibrant diversity lie a few outliers, like the circulating tumour cells, cancer stem cells, and some belonging to our immune system, which are rare. Identifying these cells that are few and far between, is a Herculean task. In a recent study, researchers from the Indian Institute of Technology Delhi, and the Indraprastha Institute of Information Technology, Delhi, have developed an algorithm to locate such rare cells based on their genes.

The fact that some cells are few and rare doesn't mean they are unimportant. In fact, many of these play critical roles in our immune responses, replace damaged cells and some are associated with cancer. Hence, identifying them becomes important to detect and treat such conditions. In this study, published in the journal Nature Communications, the researchers propose an algorithm called Finder of Rare Entities (FiRE), that identifies such rare cells by looking at their genes. The study was partially funded by the Department of Science and Technology.

Remember those days when you were asked to ‘spot the difference’ between two almost-identical images, or pick an ‘odd one out’ from a set of similar images? Well, FiRE does something similar among thousands of identical things. It looks for the gene expression profile, or the list of all the genes expressed, of all the cells and identifies the rare one based on the differences in their profile. Each type of cells expresses different genes and hence has a unique gene expression profile. The algorithm assigns a ‘rareness score’ for each of the gene expression profile it looks up. The rarer the type of the cell, the higher is this score. It then shortlists the ‘rare’ ones based on their score.

The researchers evaluated the performance of their algorithm using various datasets. The algorithm successfully identified rare cells from a test dataset containing the expression profile of 68,000 cells. Interestingly, the researchers also identified a rare cell subtype that was previously unknown, using this algorithm during an experiment with a big dataset of mouse brain cells.  These rare cells play an essential role in the development of the pituitary gland in the brain of mammals.

The new algorithm bears great potential in detecting rare cells and diseases. Although a few such algorithms exist, they are very slow. “FiRE took around 31 seconds to analyse a single cell mRNA sequencing dataset containing about 68000 expression profiles. Such unrivaled speed, combined with the ability to pinpoint the truly rare expression profiles, makes the algorithm future proof”, say the researchers.

With the growth in technologies to generate a massive amount of biological data, like the expression profile of genes from individual cells, there is a growing need for the development of tools to analyse those to retrieve the information in the datasets. Algorithms like FiRE, which are fast and efficient would greatly benefit the research community.