# CellMap: precision mapping of cellular landscape in spatial transcriptomics

## TL;DR

CellMap is a new tool that improves the precision of mapping cellular landscapes in spatial transcriptomics by integrating single-cell RNA sequencing data.

## Contribution

CellMap introduces a novel computational approach combining seed genes, random forest models, and linear assignment algorithms for high-resolution spatial mapping.

## Key findings

- CellMap achieves optimal assignment of single cells to spatial spots using co-linearity of seed genes and machine learning.
- Comprehensive benchmarking shows CellMap outperforms existing methods across multiple platforms and tissue types.

## Abstract

Integrating single-cell RNA sequencing and spatial transcriptomics is the current imperative to manually explore the landscape of cellular mixtures. Herein, we developed CellMap (https://github.com/liuhong-jia/CellMap), a computational tool that allows spatial transcriptomic spots to be resolved at single-cell resolution. CellMap combines strategies that incorporate the co-linearity of seed genes, the random forest model, and the linear assignment algorithm to achieve optimal assignment of single cells to spatial spots. Using comprehensive benchmarking across various platforms and tissue types, we demonstrated that CellMap outperforms existing methods.

Graphical Abstract