The html page source and javascript code of the online database is available online at http://betsholtzlab.org/VascularSingleCells/database.html. In order to identify enriched genes in specific brain cell type(s), the average expression for each cell types was stored in a MySQL (version 5.0.12-dev) database table and user queries were passed through a PHP (version 7.0.23) script to the MySQL database. Code availability The R code used to process the sequencing data and visualize the results is available in the Supplementary File 1 (R version 3.3.2). Data Records The information table for all the cells used in this study is Vezf1 available on Figshare (Data Citation 1). found to correspond to: pericytes, three types of vascular easy muscle cells (venous, arteriolar and arterial), microglia, two types of fibroblast-like cells, oligodendrocyte-lineage cells, six types of endothelial cells (venous, capillary, arterial and three others) and astrocytes (Fig. 2a). In the lung, we defined 17 cell clusters. Because our main objective with the lung dataset was to compare brain and lung pericytes, the annotation process of lung cells other than pericytes and endothelial Retinyl acetate cells was less extensive, but nevertheless indicated the presence of several subtypes of fibroblasts (split in four clusters) and cartilage/perichondrium-related cells (two clusters), pericytes (one cluster), vascular easy muscle cells (one cluster), and at least two distinct types of endothelial cells (split into eight clusters) (Fig. 2b). To allow the scientific community to Retinyl acetate contribute to the further annotation of these cell Retinyl acetate types by assessing their gene expression, we provide user-friendly access to our data in the form of a searchable database http://betsholtzlab.org/VascularSingleCells/database.html, in which any gene can be searched by acronym, and its expression across the analyzed cell types in brain and lung displayed as single-cell bar-plots as well as diagrams displaying average values for the expression in the different cell types (see Fig. 3a-d for an example). Open in a separate window Physique 2 Overview of the single cell data in the adult mouse brain and lung.(a) The 3,418 brain single cells were analyzed by the T-Distributed Stochastic Neighbor Embedding (splice junction reads, filtered for only uniquely mapping reads. The STAR parameters are as follows: STAR –runThreadN 1 –genomeDir mm10 –readFilesIn XXX.fastq.gz –readFilesCommand zcat –outSAMstrandField intronMotif –twopassMode Basic The expression values were computed per gene as described in Ramsk?ld et al.10, using uniquely aligned reads and correcting for the uniquely alignable positions using MULTo57(ref. 11). As QC threshold, cells with less than 100,000 reads were discarded, as well as cells that had a Spearman correlation below 0.3. Our analyses and cell type annotations were based on 3,186 brain vascular-associated cells, 1,504 lung vascular-associated cells and 250 brain Retinyl acetate astrocytes, which were obtained in parallel experiments using different reporter mice and partly different procedures to obtain the cells (see ref. 4). Therefore, in order to compare the gene expression counts across different cells, the total gene counts for each cell were normalized to 500,000. The R code used for the normalization is available in the Supplementary File 1. The R tsne packages (version 0.1.3) was applied to visualize the 2D t-SNE map and GGally packages (version 1.3.1) was used to make gene pairs plot. Cell type classification with BackSPIN As a clustering method, the BackSPIN algorithm12 was applied to classify the cells into different cell types. The BackSPIN software was downloaded from https://github.com/linnarsson-lab/BackSPIN (2015 version). BackSPIN was run with the following parameters: backspin -i input.CEF -o output.CEF -v -d 6 -g 3 -c 5 This iteratively splits the cells into six levels. After manual inspection and annotation, we defined 15 cell clusters in the brain and 17 cell clusters in the lung4. Online database construction The expression database was constructed using html and javascript. For each gene, four figures were pre-made and stored around the server for faster display (see Fig. 3a-d for an example), including: the detailed expression in each cell in the brain dataset (Fig. 3a); the average expression level in each of the 15 clusters in the brain (Fig. 3b); the detailed expression in each cell in the lung dataset (Fig. 3c) and the average expression level in each of the 17 clusters in the lung (Fig. 3d). The gene symbol auto-complete function was implemented using the jquery.autocomplete.min.js and jquery-1.9.1.min.js plugin (available from https://github.com/devbridge/jQuery-Autocomplete/). The html page source and javascript code of the online database is available online at http://betsholtzlab.org/VascularSingleCells/database.html. In order to identify enriched genes in specific brain cell type(s), the average expression for each cell types was stored in a MySQL (version 5.0.12-dev) database table and user queries were passed through a PHP (version 7.0.23) script to.