scicast

Single Cell Iterative Clustering and Significance Testing

Find futher instructions and vignette here: https://iandriver.github.io/scicast/

Requires

• Python 2.7 or 3.4+ (preferred)

• tkinter conda install tk or brew install python --with-brewed-tk

• numpy

• scipy

• scikit-learn

• matplotlib

• pandas >= 0.19.0

• seaborn >= 0.7.1

• R >3.3 -must have qgraph installed in R for qgraph_plot to work

        install.packages("qgraph")
  

• rpy2

• fastcluster

pip or brew or conda install any that are missing, prior to scicast installation.

Note on rpy2

Install or update R, and then pip install rpy2. Otherwise it won't link properly.

Installation

To install the released version, just type:

pip install scicast

Example command

Simple Command: There is now a gui (tkinter based):

          scicast -gui

will launch the gui interface:

scicast GUI window

otherwise:

          scicast -f path/to/krasnow_AT2_185_rsem_deseq_counts_norm.txt -n krasnow_example_data

Command with cell and gene files:

          scicast -f /Path/to/gene_cell_matrix -n my_gene_cell_matrix_analyis -v -cell_group sc_cell_groups.csv -gene_list sc_gene_groups.csv -depth 100 -z 0 -genes_corr Myc,Nanog,Vegf -method ward -metric seuclidean -annotate_gene_subset gene_annotation_subset.csv -add_ellipse

Example Dataset

For the soon to be vignette and for trying out scicast, an example single cell dataset is provided: krasnow_AT2_185_rsem_deseq_counts_norm.txt

From the paper: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4145853/ Raw sequencing was mapped and aligned to GRCm38 using rsem and normalized using Deseq2. Only E18.5 and Adult At2 cells are included.

Example k-means cluster:

Heatmap for k-means assignment:

Gene Correlation search:

=======

Usage

usage: scicast

  -gui        To run gui instead of command line. (default: False)

else:

              [-h] -f FILEPATH [-n BASE_NAME] [-g GENE_NUMBER] [-v]
              [-cell_group CELL_LIST_FILENAME]
              [-gene_list GENE_LIST_FILENAME] [-group_sig_test]
              [-stability TEST_CLUST_STABILITY] [-metric METRIC]
              [-method METHOD] [-depth DEPTH_OF_CLUSTERING]
              [-genes_corr GENES_CORR] [-all_sig] [-already_log2]
              [-z Z_DIRECTION] [-no_corr] [-no_heatmaps]
              [-exclude_genes EXCLUDE_GENES] [-limit_cells] [-add_ellipse]
              [-annotate_gene_subset ANNOTATE_GENE_SUBSET]
              [-annotate_cell_pca] [-color_cells COLOR_CELLS]
              [-color_genes COLOR_GENES] [-qgraph_plot QGRAPH_PLOT]
              [-sig_unique] [-kmeans_cluster_range KMEANS_CLUSTER_RANGE]
              [-kmeans_sig_test]

optional arguments:

-h, --help show this help message and exit

-f FILEPATH, -filepath FILEPATH

                    Filepath to cell-gene matrix file (default: None)

-n BASE_NAME, -name BASE_NAME

                    The base name for files that will be generated.
                    (default: scicast_analysis)

-g GENE_NUMBER, -gene_number GENE_NUMBER

                    The number of genes that will be selected from top pca
                    genes on each iteration. (default: 200)

-v, -verbose

                    Print more (default: False)

-cell_group CELL_LIST_FILENAME

                    Provide a filename with two columns: 'SampleID' and
                    'GroupID'. If GroupID is empty the SampleID list will
                    be used to restrict cells prior to analysis. (default:
                    False)

-gene_list GENE_LIST_FILENAME

                    Path to a file with a two columns with headers
                    'GeneID' and 'GroupID' (Singular format). GeneID list
                    will be used to create a new matrix file with only
                    those genes included. (default: False)

-group_sig_test

                    If cell groups are provided, perform significance
                    testing between all groups (independent of any cluster
                    groups). (default: False)

-stability TEST_CLUST_STABILITY

                    Provide a number of iterations to test how stable
                    clustering is as the number of top PCA genes changes
                    from 100-1000. Output will be clustering heatmaps for
                    each iteration a summary of changes as gene number
                    varies. (default: 0)

-metric METRIC

                    The distance metric to use. The distance function can
                    be: braycurtis, canberra, chebyshev, cityblock,
                    correlation, cosine, dice, euclidean, hamming,
                    jaccard, kulsinski, mahalanobis, matching, minkowski,
                    rogerstanimoto, russellrao, seuclidean, sokalmichener,
                    sokalsneath, sqeuclidean, yule. (default: euclidean)

-method METHOD

                    The linkage method to use. The linkage algorithm can
                    be: single, complete, average, weighted, centroid,
                    median or ward. (default: weighted)

-depth DEPTH_OF_CLUSTERING

                    The size in cell number at which sub-clustering
                    analysis will stop clustering, pca and correlation
                    analysis. (default: 20)

-genes_corr GENES_CORR

                    If you want to look for correlation on a specific gene
                    or set of genes enter them as a comma seperated list
                    i.e. 'Gapdh,Actb'. (default: )

-all_sig

                    Do significance testing on all hierarchical clustering
                    groups. The minimum number of cells in a group is set
                    by --depth. (default: False)

-already_log2

                    If the gene matrix is already log2 transformed i.e.
                    rld or vsd DESeq2, this will disable log2 transform.
                    (default: False)

-z Z_DIRECTION

                    Either 0 (rows) or 1 (columns) or None. Whether or not
                    to calculate z-scores for the rows or the columns. Z
                    scores are: z = (x - mean)/std, so values in each row
                    (column) will get the mean of the row (column)
                    subtracted, then divided by the standard deviation of
                    the row (column). This ensures that each row (column)
                    has mean of 0 and variance of 1. (default: 0)

-no_corr

                    Don't run correlation search. Default is on. (default:
                    False)

-no_heatmaps

                    Don't run heatmaps and pca. Default is will generate
                    heatmaps and pca. (default: False)

-exclude_genes EXCLUDE_GENES

                    Filepath to list of genes to be excluded from analysis
                    (at all levels of analysis). Header must be 'GeneID'.
                    (default: False)

-limit_cells

                    With cell group file, will exclude all other cells
                    from analysis (at all levels of analysis). Header must
                    be 'SampleID'. (default: False)

-add_ellipse

                    When present colored ellipses will be added to cell
                    and gene PCA plots. Must provide gene and/or cell
                    groups. (default: False)

-annotate_gene_subset ANNOTATE_GENE_SUBSET

                    Provide path or filename (if in same file) to file
                    with genes to be annotated on gene PCA. Must have
                    'GeneID' header. (default: False)

-annotate_cell_pca

                    Option will annotate cell PCA with cell names. Default
                    is off (False). (default: False)

-color_cells COLOR_CELLS

                    Follow with cell group names and pairs (matching cell
                    group names). Forces color and marker style on group.
                    Just color or color and marker can be provided.
                    Example: Group1,color,marker Group2,red,v (default:
                    False)

-color_genes COLOR_GENES

                    Can be 'same' if names are the same as cell groups.
                    Follow with gene group names and pairs (matching gene
                    group names). Forces color and marker style on group.
                    Just color or color and marker can be provided.
                    Example: Group1,color,marker Group2,red,v (default:
                    False)

-qgraph_plot QGRAPH_PLOT

                    Can be 'cell' 'gene' or 'both' will plot qgraph gene
                    and/or cell network and pca correlation network plot
                    to pdf. Requires both gene and cell groups to be
                    provided. (default: False)

-sig_unique

                    group_sig_test will by default find unique sets of
                    significant genes, so that the whole list has no
                    duplicates. This switches the top significant genes
                    amoung groups to allow repeats. (default: True)

-kmeans_cluster_range KMEANS_CLUSTER_RANGE

                    range of cluster sizes to run kmeans clustering
                    (inclusive) (default: 2,4)

-kmeans_sig_test

                    Run significance test between groups defined by kmeans
                    clusters. (default: False)