CellFinder is a freely available on-line resource for cell-based data in mammalian cells and in vitro cell lines. CellFinder was developed to provide easier access to different types of cell-based data, which include anatomical and microscopic images, protein expression and whole genome mRNA expression profiles. In this short communication, we describe the collection, processing and storage of gene expression data for the CellFinder resource. Studies involving the use of normal cells and/or production of pluripotent stem cells or their engineered cell types, were chosen from public data repositories of mRNA expression data. Samples in each study were manually curated with ontology terms, and the data was further processed to yield normalized expression data. Sample, experiment and molecular data were stored in a postgreSQL database. The resulting molecular database currently contains 1588 samples from 56 studies, covering 42 cell types in human and mouse. The molecular database is undergoing continuous development and will serve as a source of standardized, processed and expertly annotated data for CellFinder analysis tools such as MarkerTool and CompareTool.