Single-cell data for \"The phenotypic landscape of essential human genes\"

Single-cell level image-based phenotypic profiles for the publication, \"The phenotypic landscape of essential human genes.\" Due to the nature of the dataset, it is available here as chunked parquet file which altogether contains a table of >30 million individual by >1000 columns of metadata and corresponding extracted phenotype measurements. Note that all phenotype measurements are already normalized to the non-targeting sgRNA (negative control) cell population from within the same plate and well of each cell. Recommended access directions: (1) download the full dataset (~300 GB) (2) read the top-level parquet directory into python using dask dataframe to access the table metadata (pyarrow engine recommended), e.g., df = dask.dataframe.read_parquet('interphase-reclassified_cp_phenotype_normalized.parquet',engine='pyarrow')From there can bring the whole dataset into memory (if enough memory is available), can select chunks (indexed by sgRNA sequence expressed by each cell), or can perform parallelized processing of the dataset using dask or similar tools.