Spatial transcriptomic files for "An in vivo and in vitro spatiotemporal atlas of human midbrain development"

Files required to reproduce analysis from "An in vivo and in vitro spatiotemporal atlas of human midbrain development".<br>Contains:Three docker images: <code>commot_docker.tar</code>; <code>cell2location_docker.tar</code>; <code>midbrain_tissues_organoids_st_docker.tar</code>. Instructions on how to load these are provided below. <code>cell2location_docker.tar</code> is only used for pre-processing. <code>commot_docker.tar</code> and <code>midbrain_tissues_organoids_st_docker.tar</code> are used for the main analysis.<code>mlo_all_st.tar.gz</code> archive containing processed data and the code to reproduce the figures from the paper.<code>cell2location_results_*</code> (one for each sample): cell type deconvolution results used in the main analysis. <code>mlo_resolution075_Annot.rds</code>: annotated single-cell reference, used for cell2location deconvolution.Dockerfiles used to generate each docker image.<br>How to reproduce resultsThe <code>mlo_all_st.tar.gz</code> file can be downloaded and extracted and used to re-run the pipeline that generated the spatial transcriptomic figures. Once the necessary docker images have been downloaded and installed (see below), the pipeline can be run with:<code>tar -xzvf </code><code>mlo_all_st.tar.gz</code><code># From root of extracted directory (i.e. containing the `code` directory)</code><code>docker run \</code><code> --rm \</code><code> --volume /var/run/docker.sock:/var/run/docker.sock \</code><code> --volume "$(pwd):/tmp/in_mnt" \</code><code> --workdir /tmp/in_mnt/code \</code><code> --env HOST_MOUNT_PATH="$(pwd)" \</code><code> --oom-score-adj -1000 \</code><code> midbrain_tissues_organoids_docker \</code><code> R -e 'rmarkdown::render("mlo_analysis_code.Rmd")'</code><br><br>The above code should exactly reproduce the analysis in the publication. If it does not, please let us know.Note that <code>mlo_all_st.tar.gz </code>already contains generated figures and data. These are included so that the target is obvious to the user: re-running the pipeline will replace these files with regenerated versions.<br>Docker images<br>We used Docker to ensure reproducibility. The three images used to run the analysis can be loaded into docker with:<code>docker load &lt; IMAGE_NAME</code>The <code>cell2location</code> and <code>COMMOT</code> images were generated using the installation instructions of the packages they contain. The `midbrain_tissues_organoids_st_docker` is more complex since it contains all other software necessary to run the rest of the analysis. It is based on the docker image generated by the `Seurat` developers, with additional R packages installed. The versions of these R packages are stored in `renv.lock`, generated by the `renv` package. The Dockerfiles for each image are in this repository.<br>Running <code>Cell2Location</code><br>Before the main analysis pipeline, we run <code>cell2location</code> to estimate the cell type proportions on each spot of the Visium slides. We do so with<br><code>sh run_entire_c2l_pipeline.sh</code><br>You probably don't want to re-run this step since it is time-consuming.<br>