pycyto 0.0.1__tar.gz

pycyto-0.0.1/PKG-INFO

@@ -0,0 +1,30 @@
+Metadata-Version: 2.1
+Name: pycyto
+Version: 0.0.1
+Summary: pycyto is designed to analyze Stereo-seq data.
+Home-page: https://github.com/mgy520/pycyto
+Author: Mao Guangyao
+License: MIT License
+Requires-Python: ==3.8.18
+Requires-Dist: rpy2==3.5.15
+Requires-Dist: anndata2ri==1.3.1
+Requires-Dist: numpy==1.24.4
+Requires-Dist: diffxpy==0.7.4
+Requires-Dist: scipy==1.10.1
+Requires-Dist: pandas==2.0.3
+Requires-Dist: anndata==0.9.2
+Requires-Dist: shapely==2.0.3
+Requires-Dist: dask==2023.5.0
+Requires-Dist: numba==0.58.1
+Requires-Dist: pca==2.0.5
+Requires-Dist: openTSNE==1.0.1
+Requires-Dist: igraph==0.11.4
+Requires-Dist: sinfonia==0.0.3
+Requires-Dist: tangram==0.7.0
+Requires-Dist: scanorama==1.7.4
+Requires-Dist: matplotlib==3.7.4
+Requires-Dist: seaborn==0.13.2
+Requires-Dist: adjustText==1.1.1
+Requires-Dist: scikit-learn==1.3.2
+Requires-Dist: scikit-misc==0.2.0
+Requires-Dist: scikit-image==0.21.0

pycyto-0.0.1/pycyto/Findmarkers.py

@@ -0,0 +1,74 @@
+import diffxpy.api as de
+import numpy as np
+from sklearn.utils.sparsefuncs import mean_variance_axis
+from scipy.sparse import issparse
+
+
+def findallmarkers(cdata, cluster_key, min_pct=0.1, only_pos=True):
+    """
+        Find marker genes for each cluster in cdata.
+
+        Parameters:
+            cdata (Anndata): An Anndata object.
+            cluster_key (str): The key representing the cell cluster annotations in the observation metadata of the Anndata object.
+            min_pct (float, optional): The minimum percentage threshold for filtering sparse genes. Defaults to 0.1.
+            only_pos (bool, optional): Whether to consider only genes with positive log2 fold change. Defaults to True.
+
+        Returns:
+            cdata (Anndata): An Anndata object with marker genes identified for each cell cluster. The marker genes and related statistical information are stored in the `uns` attribute of the Anndata object under the key 'markers_all'.
+        """
+    clusters = cdata.obs[cluster_key].unique()
+    cluster_markers = {}
+    for cluster in clusters:
+        cdata.obs['group'] = ['other' if x == cluster else 'current' for x in cdata.obs[cluster_key]]
+        cluster_cells = cdata[cdata.obs['group'] == 'current']
+        other_cells = cdata[cdata.obs['group'] == 'other']
+        pct1 = np.mean(cluster_cells.X > 0, axis=0).A1 if issparse(cluster_cells.X) else np.mean(cluster_cells.X > 0, axis=0)
+        pct2 = np.mean(other_cells.X > 0, axis=0).A1 if issparse(other_cells.X) else np.mean(other_cells.X > 0, axis=0)
+        if issparse(cdata.X):
+            mean1, _ = mean_variance_axis(axis=0, X=cluster_cells.X)
+            mean2, _ = mean_variance_axis(axis=0, X=other_cells.X)
+        else:
+            mean1 = cluster_cells.X.mean(axis=0)
+            mean2 = other_cells.X.mean(axis=0)
+
+        selected_genes = (pct1 >= min_pct) | (pct2 >= min_pct)
+        np.float = float
+        test_result = de.test.t_test(
+            data=cdata[:, selected_genes],
+            grouping=cdata.obs["group"],
+            is_logged=True
+        )
+        result = test_result.summary()
+        result['pct1'] = pct1[selected_genes]
+        result['pct2'] = pct2[selected_genes]
+        if only_pos:
+            result = result[result['log2fc'] > 0]
+
+        cluster_markers[cluster] = result
+    cdata.uns['markers_all'] = cluster_markers
+
+    return cdata
+
+
+def find_markers_between_groups(cdata, cluster_key, group1, group2, min_pct=0.1, only_pos=True):
+    """
+        Find marker genes between two specified groups in cdata.
+
+        Parameters:
+            cdata (Anndata): An Anndata object containing the single-cell data.
+            cluster_key (str): The key representing the cell cluster annotations in the observation metadata of the Anndata object.
+            group1 (str): The name of the first group for comparison.
+            group2 (str): The name of the second group for comparison.
+            min_pct (float, optional): The minimum percentage threshold for filtering sparse genes. Defaults to 0.1.
+            only_pos (bool, optional): Whether to consider only genes with positive log2 fold change. Defaults to True.
+
+        Returns:
+            cdata(Anndata): An Anndata object with marker genes identified between the specified groups. The marker genes and related statistical information are stored in the `uns` attribute of the Anndata object under a key formatted as 'DGE_group1_group2'.
+        """
+    tmp = cdata[(cdata.obs[cluster_key] == group1) | (cdata.obs[cluster_key] == group2)].copy()
+    tmp = findallmarkers(cdata=tmp, cluster_key=cluster_key, min_pct=min_pct, only_pos=only_pos)
+    key_name = f"DGE_{group1}_{group2}"
+    cdata.uns[key_name] = tmp.uns['markers_all']
+
+    return cdata

pycyto-0.0.1/pycyto/QC.py

@@ -0,0 +1,333 @@
+import numpy as np
+import matplotlib.pyplot as plt
+import seaborn as sns
+from pycyto import function
+
+def markers_detected_by_diff_ngenes_scatter(adata, markers, image_width=6, image_height=6, xlabel='n_genes',
+                                            ylabel='Markers detection rate',title=None,o=None, **kwargs):
+    """
+        Create a scatter plot to visualize the relationship between the number of genes (n_genes) and the detection rate of markers.
+
+        Parameters:
+            adata (AnnData):An AnnData object.
+            markers (list): A list of marker gene names.
+            image_width (int, optional): Width of the generated image. Default is 6.
+            image_height (int, optional): Height of the generated image. Default is 6.
+            xlabel (str, optional): Label for the X-axis. Default is 'n_genes'.
+            ylabel (str, optional): Label for the Y-axis. Default is 'Markers detection rate'.
+            title (str, optional): Title of the plot. If not provided, no title is displayed.
+            o (str, optional): Path and filename for saving the image. If not provided, the image will not be saved.
+            **kwargs (typing.Any): Additional keyword arguments to be passed to the plt.scatter function.
+        """
+    adata.obs['n_genes'] = (adata.X > 0).sum(axis=1)
+    p = []
+    for i in range(len(adata)):
+        genes = adata.var_names[adata.X[i].nonzero()[1]]
+        intersect = set(genes).intersection(set(markers))
+        p.append(len(intersect) / len(set(markers)))
+
+    adata.obs["Markers detection rate"] = p
+    plt.figure(figsize=(image_width, image_height))
+    plt.scatter(adata.obs['n_genes'], adata.obs['Markers detection rate'], **kwargs)
+    plt.xlabel(xlabel)
+    plt.ylabel(ylabel)
+    plt.title(title)
+
+    if o is not None and isinstance(o, str):
+        plt.savefig(o)
+
+    plt.show()
+
+def markers_detected_by_diff_ngenes_bar(adata, markers, bin=50, image_width=12, image_height=6, xlabel='n_genes',
+                                        ylabel='Markers detection rate', title=None, o=None, **kwargs):
+    """
+        Create a bar plot to visualize the relationship between the number of genes (n_genes) and the detection rate of markers.
+
+        Parameters:
+            adata (AnnData): An AnnData object.
+            markers (list): A list of marker gene names.
+            bin (int, optional): The bin size for grouping n_genes values. Default is 50.
+            image_width (int, optional): Width of the generated image. Default is 12.
+            image_height (int, optional): Height of the generated image. Default is 6.
+            xlabel (str, optional): Label for the X-axis. Default is 'n_genes'.
+            ylabel (str, optional): Label for the Y-axis. Default is 'Markers detection rate'.
+            title (str, optional): Title of the plot. If not provided, no title is displayed.
+            o (str, optional): Path and filename for saving the image. If not provided, the image will not be saved.
+            **kwargs (typing.Any): Additional keyword arguments to be passed to the plt.bar function.
+        """
+    adata.obs['n_genes'] = (adata.X > 0).sum(axis=1)
+    bins = np.arange(0, np.ceil(adata.obs['n_genes'].max() / bin) * bin, bin).astype(int)
+    adata.obs['n_genes_subset'] = np.digitize(adata.obs['n_genes'], bins)
+
+    counts = []
+    for i in range(len(adata)):
+        genes = adata.var_names[adata.X[i].nonzero()[1]]
+        intersect = set(genes).intersection(set(markers))
+        counts.append(len(intersect))
+
+    adata.obs["Markers intersect counts"] = counts
+
+    subset_counts = []
+    for i in range(len(bins) - 1):
+        subset = adata[adata.obs['n_genes_subset'] == i + 1]
+        n = 0
+        for cell in subset.obs_names:
+            n += subset.obs.loc[cell, 'Markers intersect counts']
+        if len(subset.obs_names) > 0:
+            n = n / len(subset.obs_names)
+            p = n / len(set(markers))
+            subset_counts.append(p)
+        else:
+            subset_counts.append(0)
+
+    plt.figure(figsize=(image_width, image_height))
+    plt.bar([f"{bins[i]}-{bins[i+1]}" for i in range(len(bins) - 1)], subset_counts, color='#898989',**kwargs)
+    plt.xticks(rotation=90)
+    plt.ylabel(ylabel)
+    plt.xlabel(xlabel)
+    plt.title(title)
+
+    if o is not None and isinstance(o, str):
+        plt.savefig(o)
+
+    plt.show()
+
+def detect_markers_by_filter_cells(adata, markers, min_genes_list=list(range(0, 1400, 50)), min_cells_list=[3],
+                                   image_width=5, image_height=5, xlabel='min_genes', ylabel='Percentage',
+                                   title='Percentage of markers remaining',o=None, **kwargs):
+    """
+        Create a line plot that shows how many of the marker genes are still present in the adata after filtering,
+        as you change the min_genes and the min_cells used for filtering.(The proportion of marker genes in markers)
+
+        Parameters:
+            adata (AnnData): An AnnData object.
+            markers (list): A list of marker gene names.
+            min_genes_list (list, optional): A list of min_genes to consider. Default list(range(0, 1400, 50)).
+            min_cells_list (list, optional): A list of min_cells for gene filtering. Default [3].
+            image_width (int, optional): Width of the generated image. Default is 5.
+            image_height (int, optional): Height of the generated image. Default is 5.
+            xlabel (str, optional): Label for the X-axis. Default is 'min_genes'.
+            ylabel (str, optional): Label for the Y-axis. Default is 'Percentage'.
+            title (str, optional): Title of the plot. Default is 'Percentage of markers remaining'.
+            o (str, optional): Path and filename for saving the image. If not provided, the image will not be saved.
+            **kwargs (typing.Any): Additional keyword arguments to be passed to the plt.plot function.
+        """
+    marker_overlap = {frac: [] for frac in min_cells_list}
+    for frac in min_cells_list:
+        for min_genes in min_genes_list:
+            adata_filtered = adata.copy()
+            adata_filtered = function.filter_cells(adata_filtered, min_genes=min_genes)
+            adata_filtered = function.filter_genes(adata_filtered, min_cells=frac)
+            remaining_genes = adata_filtered.var_names.tolist()
+            overlap_genes = set(remaining_genes).intersection(set(markers))
+            marker_overlap_percent = len(overlap_genes) / len(set(markers))
+            marker_overlap[frac].append(marker_overlap_percent)
+
+    plt.figure(figsize=(image_width, image_height))
+    for frac in min_cells_list:
+        plt.plot(min_genes_list, marker_overlap[frac], label=f'min_cells={frac}', **kwargs)
+    plt.xlabel(xlabel)
+    plt.ylabel(ylabel)
+    plt.title(title)
+    plt.legend(bbox_to_anchor=(1.05, 0.5), loc='center left')
+    if o is not None and isinstance(o, str):
+        plt.savefig(o)
+    plt.show()
+
+def detect_markers_by_filter_cells2(adata, markers, min_genes_list=list(range(0, 1400, 50)), min_cells_list=[3],
+                                   image_width=5, image_height=5, xlabel='min_genes', ylabel='Percentage',
+                                   title='Percentage of markers detected',o=None, **kwargs):
+    """
+        Create a line plot that shows how many of the marker genes are still present in the adata after filtering,
+        as you change the min_genes and the min_cells used for filtering.(The proportion of marker genes in adata)
+
+        Parameters:
+            adata (AnnData): An AnnData object.
+            markers (list): A list of marker gene names.
+            min_genes_list (list, optional): List of minimum gene counts for filtering. Default list(range(0, 1400, 50)).
+            min_cells_list (list, optional): List of minimum cell counts for filtering. Default is [3].
+            image_width (int, optional): Width of the generated image. Default is 5.
+            image_height (int, optional): Height of the generated image. Default is 5.
+            xlabel (str, optional): Label for the X-axis. Default is 'min_genes'.
+            ylabel (str, optional): Label for the Y-axis. Default is 'Percentage'.
+            title (str, optional): Title of the plot. Default is 'Percentage of markers detected'.
+            o (str, optional): Path and filename for saving the image. If not provided, the image will not be saved.
+            **kwargs (typing.Any): Additional keyword arguments to be passed to the plt.plot function.
+        """
+    ...
+    gene_overlap = {frac: [] for frac in min_cells_list}
+    for frac in min_cells_list:
+        for min_genes in min_genes_list:
+            adata_filtered = adata.copy()
+            adata_filtered = function.filter_cells(adata_filtered, min_genes=min_genes)
+            adata_filtered = function.filter_genes(adata_filtered, min_cells=frac)
+            remaining_genes = adata_filtered.var_names.tolist()
+            overlap_genes = set(remaining_genes).intersection(set(markers))
+            overlap_percent = len(overlap_genes) / len(remaining_genes)
+            gene_overlap[frac].append(overlap_percent)
+
+    plt.figure(figsize=(image_width, image_height))
+    for frac in min_cells_list:
+        plt.plot(min_genes_list, gene_overlap[frac], label=f'min_cells={frac}', **kwargs)
+    plt.xlabel(xlabel)
+    plt.ylabel(ylabel)
+    plt.title(title)
+    plt.legend(bbox_to_anchor=(1.05, 0.5), loc='center left')
+    if o is not None and isinstance(o, str):
+        plt.savefig(o)
+    plt.show()
+
+def cell_numbers_by_min_genes(adata, min_genes_list=list(range(0, 1400, 50)),
+                                   image_width=5, image_height=5, xlabel='min_genes', ylabel='cell number',
+                                   title=None,o=None, **kwargs):
+    """
+        Create a line plot reflecting the number of cells remaining after filtering with different min_genes
+
+        Parameters:
+            adata (AnnData): An AnnData object.
+            min_genes_list (list, optional): List of minimum gene counts used for filtering cells. Default list(range(0, 1400, 50)).
+            image_width (int, optional): Width of the generated plot image. Default is 5.
+            image_height (int, optional): Height of the generated plot image. Default is 5.
+            xlabel (str, optional): Label for the X-axis of the plot. Default is 'min_genes'.
+            ylabel (str, optional): Label for the Y-axis of the plot. Default is 'cell number'.
+            title (str, optional): Title of the plot. If not provided, no title will be displayed.
+            o (str, optional): Path and filename for saving the plot image. If not provided, the image will not be saved.
+            **kwargs (typing.Any): Additional keyword arguments to be passed to the plt.plot function.
+        """
+    num_cells = []
+    for min_gene in min_genes_list:
+        adata_filtered = adata.copy()
+        adata_filtered = function.filter_cells(adata_filtered, min_genes=min_gene)
+        num_cells.append(adata_filtered.shape[0])
+
+    plt.figure(figsize=(image_width, image_height))
+    plt.plot(min_genes_list, num_cells, **kwargs)
+    plt.xlabel(xlabel)
+    plt.ylabel(ylabel)
+    plt.title(title)
+    if o is not None and isinstance(o, str):
+        plt.savefig(o)
+    plt.show()
+
+def valid_cells_per_slice(adata, min_genes_list=list(range(0, 800, 100)), image_width=5, image_height=5, xlabel='min_genes',
+                          slice='id', library='current object', median_n_genes=[],ylabel='Percentage',title='Percentage of markers detected',
+                          o=None, **kwargs):
+    """
+        Create line plots to illustrate the relationship between the library median n_genes and the percentage of valid cells per slice.
+
+        Parameters:
+            adata (AnnData): An AnnData object.
+            min_genes_list (list, optional): A list of minimum gene counts for cell filtering. Default is a list from 0 to 800 with a step of 100.
+            image_width (int, optional): Width of the generated image. Default is 5.
+            image_height (int, optional): Height of the generated image. Default is 5.
+            xlabel (str, optional): Label for the X-axis. Default is 'min_genes'.
+            slice (str): Column in the observation data used for grouping slices and analyzing the percentage of valid cells.
+            library (str): Source of data for calculating the median number of genes per slice. Default is 'current object', using the observation data of the current object. Alternatively, provide a list 'median_n_genes' as reference.
+            median_n_genes (list, optional): If 'library' is not 'current object', provide a list of Library median gene counts as reference.
+            ylabel (str, optional): Label for the Y-axis. Default is 'Percentage'.
+            title (str, optional): Title of the plot. Default is 'Percentage of markers detected'.
+            o (str, optional): Path and filename for saving the image. If not provided, the image will not be saved.
+            **kwargs (typing.Any): Additional keyword arguments to be passed to the sns.regplot function.
+        """
+    adata.obs['n_genes'] = (adata.X > 0).sum(axis=1)
+    if library == 'current object':
+        library_median_n_genes = adata.obs.groupby(slice)['n_genes'].median().tolist()
+    elif median_n_genes is not None:
+        if not isinstance(median_n_genes, list):
+            raise ValueError("'median_n_genes' must be a list.")
+        library_median_n_genes = median_n_genes
+    else:
+        raise ValueError("If 'library' is not 'current object', 'median_n_genes' must be provided as a list.")
+
+    plt.figure(figsize=(image_width, image_height))
+    #colors = ['red', 'blue', 'green', 'orange', 'purple', 'gray', 'cyan', 'magenta']
+    s1 = adata.obs.groupby(slice).size()
+    s2_list = []
+    for i, min_genes in enumerate(min_genes_list):
+        adata_filtered = adata.copy()
+        adata_filtered = function.filter_cells(adata_filtered, min_genes=min_genes)
+        s2 = adata_filtered.obs.groupby(slice).size()
+        s2_list.append(s2)
+        p = s2 / s1
+        index_mapping = list(range(len(s1.index)))
+        sns.regplot(x=[library_median_n_genes[index_mapping[j]] for j in range(len(index_mapping))], y=p,
+                    scatter_kws={'s': 2},label=f'min_genes={min_genes}', **kwargs)
+        #sns.regplot(x=[library_median_n_genes[index_mapping[j]] for j in range(len(index_mapping))], y=p,scatter_kws={'s': 2}, label=f'min_genes={min_genes}', color=colors[i])
+
+    plt.xlabel(xlabel)
+    plt.ylabel(ylabel)
+    plt.legend(fontsize=6)
+    plt.title(title)
+    if o is not None and isinstance(o, str):
+        plt.savefig(o)
+    plt.show()
+
+def additional_seq_vs_seq(adata1, adata2, min_genes_list=list(range(0, 800, 100)), image_width=5, image_height=5, xlabel='min_genes',
+                          slice='id', min_cells=3, ylabel='valid cells P.',title='sequencing vs additional sequencing',slice_id=None,
+                          color='blue', s=15, o=None):
+    """
+        Create a scatter plot to compare the percentage of valid cells in sequencing data and additional sequencing data.
+
+        Parameters:
+            adata1 (AnnData): An AnnData object representing the sequencing data.
+            adata2 (AnnData): An AnnData object representing the additional sequencing data.
+            min_genes_list (list, optional): A list of minimum gene counts for cell filtering. Default is a list from 0 to 800 with a step of 100.
+            image_width (int, optional): Width of the generated image. Default is 5.
+            image_height (int, optional): Height of the generated image. Default is 5.
+            xlabel (str, optional): Label for the X-axis. Default is 'min_genes'.
+            slice (str): Column in the observation data used for grouping slices.
+            min_cells (int): min_cells to filter genes.
+            ylabel (str, optional): Label for the Y-axis. Default is 'valid cells P.'.
+            title (str, optional): Title of the plot. Default is 'sequencing vs additional sequencing'.
+            slice_id (str): The ID of the slice to analyze.
+            color (str): Color for the scatter plot markers. Default is 'blue'.
+            s (int): Size of the scatter plot markers. Default is 15.
+            o (str, optional): Path and filename for saving the image. If not provided, the image will not be saved.
+        """
+    adata1 = adata1[(adata1.obs[slice] == slice_id),]
+    adata2 = adata2[(adata2.obs[slice] == slice_id),]
+    s1_adata1 = adata1.obs.groupby(slice).size()
+    s1_adata2 = adata2.obs.groupby(slice).size()
+    p_adata1_list = []
+    p_adata2_list = []
+    for min_genes in min_genes_list:
+        adata1_filtered = adata1.copy()
+        adata2_filtered = adata2.copy()
+        adata1_filtered = function.filter_cells(adata1_filtered, min_genes=min_genes)
+        adata1_filtered = function.filter_genes(adata1_filtered, min_cells=min_cells)
+        adata2_filtered = function.filter_cells(adata2_filtered, min_genes=min_genes)
+        adata2_filtered = function.filter_genes(adata2_filtered, min_cells=min_cells)
+
+        s2_adata1 = adata1_filtered.obs.groupby(slice).size()
+        s2_adata2 = adata2_filtered.obs.groupby(slice).size()
+
+        p_adata1 = s2_adata1 / s1_adata1
+        p_adata2 = s2_adata2 / s1_adata2
+        p_adata1_list.append(p_adata1)
+        p_adata2_list.append(p_adata2)
+
+    plt.figure(figsize=(image_width, image_height))
+    if slice_id is None:
+        raise ValueError("Please provide a valid slice_id.")
+
+    p_adata1 = [p_adata1_list[i][slice_id] for i in range(len(min_genes_list))]
+    p_adata2 = [p_adata2_list[i][slice_id] for i in range(len(min_genes_list))]
+
+    plt.scatter(min_genes_list, p_adata1, label=f'{slice_id}', marker='o',color=color,s=s)
+    plt.scatter(min_genes_list, p_adata2, label=f'{slice_id}', marker='^',color=color,s=s)
+
+    plt.xlabel(xlabel)
+    plt.ylabel(ylabel)
+    plt.title(title)
+
+    legend_elements = [
+        plt.Line2D([0], [0], marker='o', color='w', label=f'{slice_id}', markersize=8)]
+    legend_elements.append(
+        plt.Line2D([0], [0], marker='o', color='w', label='sequencing', markerfacecolor=color, markersize=8))
+    legend_elements.append(
+        plt.Line2D([0], [0], marker='^', color='w', label='additional sequencing', markerfacecolor=color, markersize=8))
+
+    plt.legend(handles=legend_elements,loc='center left', bbox_to_anchor=(1, 0.5))
+    if o is not None and isinstance(o, str):
+        plt.savefig(o)
+    plt.show()

pycyto-0.0.1/pycyto/__init__.py

@@ -0,0 +1,8 @@
+from .plot import *
+from .function import *
+from .format import *
+from .QC import *
+from .labeltransfer import *
+from .cluster import *
+from .Findmarkers import *
+from .spatial_metacell import *

pycyto-0.0.1/pycyto/cluster.py

@@ -0,0 +1,46 @@
+import anndata2ri
+from rpy2.robjects import r
+import numpy as np
+
+anndata2ri.activate()
+
+def cluster(cdata, K_set, neighborhood_size, n_PCs):
+    """
+        Stereo-seq data spatial clustering.
+
+        Parameters:
+            cdata (Anndata): Anndata object containing the single-cell data.
+            K_set (list): List of integers specifying the range of cluster numbers to consider.
+            neighborhood_size (int): Size of the neighborhood for computing adjacency matrix.
+            n_PCs (int): Number of principal components to use.
+
+        Returns:
+            cdata (Anndata): An Anndata object with cluster labels assigned to observations.
+        """
+    r.assign("cdata", cdata)
+    r.assign("K_set", K_set)
+    r.assign("neighborhood_size", neighborhood_size)
+    r.assign("num_pca_components", n_PCs)
+    r('''
+    cdata <- as(cdata, "SingleCellExperiment")
+    library(SC.MEB)
+    library(SingleCellExperiment)
+    pos = as.matrix(colData(cdata)[,c("row","col")])
+    Adj_sp = getneighborhood_fast(pos, neighborhood_size)
+    y = reducedDim(cdata, "PCA")[,1:n_PCs]
+    set.seed(114)
+    beta_grid = seq(0,4,0.2)
+    parallel=TRUE
+    num_core = 3
+    PX = TRUE
+    maxIter_ICM = 10
+    maxIter = 50
+    fit = SC.MEB(y, Adj_sp, beta_grid = beta_grid, K_set= K_set, parallel=parallel, num_core = num_core, PX = PX, maxIter_ICM=maxIter_ICM, maxIter=maxIter)
+    selectKPlot(fit, K_set = K_set, criterion = "BIC")
+    out = selectK(fit, K_set = K_set, criterion = "BIC")
+    ''')
+    out = r['out']
+    best_K_label_flat = np.array(out.rx2('best_K_label')).flatten()
+    cdata.obs['cluster'] = best_K_label_flat.astype(str)
+
+    return cdata