exprmat 0.1.15__tar.gz → 0.1.16__tar.gz

{exprmat-0.1.15 → exprmat-0.1.16}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: exprmat
-Version: 0.1.15
+Version: 0.1.16
 Summary: routines to process expression matrices
 Home-page: https://github.com/yang-z-03/exprmat
 Author: Zheng Yang

{exprmat-0.1.15 → exprmat-0.1.16}/exprmat/__init__.py

@@ -4,4 +4,4 @@ import anndata as ad
 
 mu.set_options(pull_on_update = False)
 
-def version(): return 'exprmat 0.1.15-1 [dev]'
+def version(): return 'exprmat 0.1.16-1 [dev]'

{exprmat-0.1.15 → exprmat-0.1.16}/exprmat/dynamics/__init__.py

@@ -48,4 +48,43 @@ def run_velocity(
     terminal_states(adata, vkey = velocity_key, neighbor_key = neighbor_key, **kwargs_terminal_state)
     velocity_pseudotime(adata, vkey = velocity_key, neighbor_key = neighbor_key, **kwargs_pseudotime)
     velocity_confidence(adata, vkey = velocity_key, neighbor_key = neighbor_key)
-    
+
+
+def clean_velocity_results(adata, vkey = 'velocity'):
+    
+    for x in list(adata.obs.keys()):
+        if x in [
+            'root.cells', 
+            'endpoints', 
+            'n.umi.spliced', 
+            'n.umi.unspliced', 
+            'velocity.pseudotime'
+        ]: del adata.obs[x]
+        elif x.startswith(f'{vkey}.'): del adata.obs[x]
+    
+    for x in list(adata.var.keys()):
+        if x.startswith(f'{vkey}.'): del adata.var[x]
+        elif x in ['gene.count.corr']: del adata.var[x]
+
+    for x in list(adata.layers.keys()):
+        if x in ['ms', 'mu', vkey, f'variance.{vkey}']: del adata.layers[x]
+        elif ('spliced.counts' in adata.layers.keys()) and x == 'spliced':
+            del adata.layers['spliced']
+            adata.layers['spliced'] = adata.layers['spliced.counts'].copy()
+            del adata.layers['spliced.counts']
+        elif ('unspliced.counts' in adata.layers.keys()) and x == 'unspliced':
+            del adata.layers['unspliced']
+            adata.layers['unspliced'] = adata.layers['unspliced.counts'].copy()
+            del adata.layers['unspliced.counts']
+    
+    for x in list(adata.obsm.keys()):
+        if x.startswith(f'{vkey}.'): del adata.obsm[x]
+        elif x.startswith('vdiff.'): del adata.obsm[x]
+        elif x == 'vdiff': del adata.obsm[x]
+
+    for x in list(adata.uns.keys()):
+        if x in [
+            f'{vkey}.graph',
+            f'{vkey}.graph.neg',
+            f'{vkey}.params'
+        ]: del adata.uns[x]

{exprmat-0.1.15 → exprmat-0.1.16}/exprmat/plotting/expression.py

@@ -148,4 +148,73 @@ def matrix(
     ax.set_xticks(range(len(var)), labels = var, rotation = 90, ha = "right", va = 'center', rotation_mode = "anchor")
     ax.set_yticks(range(len(obs)), labels = obs)
 
+    return fig
+
+
+def volcano(
+    adata, de_slot = 'deg', label = [],
+    show_all = False, min_pct = 0, max_pct_reference = 1, 
+    min_lfc = -25, max_lfc = 25, remove_zero_pval = False,
+    highlight_min_logp = 5, highlight_min_lfc = 1.5,
+    xlim = 5, ylim = 100,
+    figsize = (3, 3), dpi = 100
+):
+
+    import numpy as np
+    import matplotlib.pyplot as plt
+    from adjustText import adjust_text
+    from exprmat.reader.experiment import experiment
+
+    pvals = []
+    fc = []
+    names = []
+
+    plots = experiment.rna_get_markers(
+        adata, de_slot = de_slot,
+        min_pct = min_pct, max_pct_reference = max_pct_reference, 
+        min_lfc = min_lfc, max_lfc = max_lfc, 
+        remove_zero_pval = remove_zero_pval
+    )
+    
+    pvals = plots['log10.q'].tolist()
+    fc = plots['lfc'].tolist()
+    names = plots['gene'].tolist()
+
+    fig, axes = plt.subplots(1, 1, figsize = figsize, dpi = dpi)
+    axes.scatter(fc, pvals, c = 'k', s = 4)
+
+    sig_pvals = []
+    sig_fc = []
+    sig_names = []
+
+    for p, f, nm in zip(pvals, fc, names):
+        if nm in label:
+            sig_pvals += [p]
+            sig_fc += [f]
+            sig_names += [nm]
+
+        if (p > highlight_min_logp and abs(f) >= highlight_min_lfc) and show_all:
+            sig_pvals += [p]
+            sig_fc += [f]
+            sig_names += [nm]
+
+    texts = []
+    for p, f, nm in zip(sig_pvals, sig_fc, sig_names):
+        texts += [axes.text(f, p, nm, va = 'top', ha = 'center')]
+
+    adjust_text(texts)
+    axes.scatter(sig_fc, sig_pvals, c = 'r', s = 10)
+    axes.grid(False)
+
+    xlim = min(np.abs(np.array(fc)).max(), xlim)
+    ylim = min(ylim, np.max(np.array(pvals)))
+    axes.set_xlim([-xlim, xlim])
+    axes.set_ylim([-ylim / 20, ylim])
+
+    axes.vlines(x = [0], ymin = [0], ymax = [ylim], linestyles = 'dashed', color = 'gray', linewidth = 0.6)
+    axes.set_yticks([])
+    axes.set_xlabel('Log-2 fold changes')
+    for pos in ['right', 'top', 'left']:
+        axes.spines[pos].set_visible(False)
+        
     return fig

{exprmat-0.1.15 → exprmat-0.1.16}/exprmat/plotting/gse.py

@@ -504,6 +504,7 @@ class dotplot(object):
         y_order: Optional[List[str]] = None,
         thresh: float = 0.05,
         n_terms: int = 10,
+        terms = None,
         title: str = "",
         ax: Optional[plt.Axes] = None,
         figsize: Tuple[float, float] = (6, 5.5),
@@ -525,6 +526,7 @@ class dotplot(object):
         self.scale = dot_scale
         self.title = title
         self.n_terms = n_terms
+        self.terms = terms
         self.thresh = thresh
         self.data = self.process(df)
 
@@ -569,19 +571,26 @@ class dotplot(object):
             self.cbar_title = f'-log({self.colname})'
             self.colname = "p_inv"
 
+        if self.terms is not None:
+            if 'term' in df.columns:
+                df = df.loc[[x in self.terms for x in df['term']], :].copy()
+            elif 'name' in df.columns:
+                df = df.loc[[x in self.terms for x in df['name']], :].copy()
+            df = df.sort_values(by = self.colname)
+        
         # get top terms; sort ascending
-        if ((self.x is not None)
+        elif ((self.x is not None)
             and (self.x in df.columns)
             and (not all(df[self.x].map(self.isfloat)))
         ):
 
             df = ( # if x is numeric column, get top term of each group
                 df.groupby(self.x)
-                .apply(lambda _x: _x.sort_values(by=self.colname).tail(self.n_terms))
-                .reset_index(drop=True)
+                  .apply(lambda _x: _x.sort_values(by = self.colname).tail(self.n_terms))
+                  .reset_index(drop=True)
             )
 
-        else: df = df.sort_values(by=self.colname).tail(self.n_terms)  # acending
+        else: df = df.sort_values(by = self.colname).tail(self.n_terms)  # acending
         
         # get scatter area
         if df.columns.isin(["overlap", "tag"]).any():
@@ -794,6 +803,7 @@ def gsea_dotplot(
     title: str = "",
     cutoff: float = 0.05,
     top_term: int = 10,
+    terms = None,
     size: float = 5,
     ax: Optional[plt.Axes] = None,
     figsize: Tuple[float, float] = (4, 6),
@@ -816,6 +826,7 @@ def gsea_dotplot(
         title = title,
         thresh = cutoff,
         n_terms = int(top_term),
+        terms = terms,
         dot_scale = size,
         ax = ax,
         figsize = figsize,
@@ -848,6 +859,7 @@ def opa_dotplot(
     title: str = "",
     cutoff: float = 0.05,
     top_term: int = 10,
+    terms = None,
     size: float = 5,
     ax: Optional[plt.Axes] = None,
     figsize: Tuple[float, float] = (4, 6),
@@ -870,6 +882,7 @@ def opa_dotplot(
         title = title,
         thresh = cutoff,
         n_terms = int(top_term),
+        terms = terms,
         dot_scale = size,
         ax = ax,
         figsize = figsize,

{exprmat-0.1.15 → exprmat-0.1.16}/exprmat/preprocessing/qc.py

@@ -123,10 +123,14 @@ def rna_qc(
     adata.obs['pct.ribo'] = adata.obs['n.ribo'] / adata.obs['n.umi']
 
     # detect outliers
+    max_genes = None
     if outlier_mode == 'mads':
         umi_lower, umi_upper = mads(adata.obs['n.umi'].to_numpy(), nmads = outlier_n)
     elif outlier_mode == 'tukey':
         umi_lower, umi_upper = tukey(adata.obs['n.umi'].to_numpy(), n = outlier_n)
+    elif isinstance(outlier_mode, int):
+        umi_lower, umi_upper = 200, 100000
+        max_genes = outlier_mode
     else: umi_lower, umi_upper = 200, 100000
 
     f_obs = \
@@ -134,6 +138,8 @@ def rna_qc(
         (adata.obs['pct.mito'] < mt_percent) & \
         (adata.obs['n.umi'] >= umi_lower) & \
         (adata.obs['n.genes'] >= min_genes)
+    if max_genes is not None:
+        f_obs = f_obs & (adata.obs['n.genes'] <= max_genes)
     
     if ribo_percent is not None:
         f_obs = f_obs & (adata.obs['pct.ribo'] < ribo_percent)

{exprmat-0.1.15 → exprmat-0.1.16}/exprmat/reader/experiment.py

@@ -34,7 +34,12 @@ from exprmat.ansi import warning, info, error, red, green
 
 class experiment:
     
-    def __init__(self, meta : metadata, mudata = None, modalities = {}, dump = '.', subset = None):
+    def __init__(
+        self, meta : metadata, 
+        eccentric = None, 
+        mudata = None, modalities = {}, 
+        dump = '.', subset = None
+    ):
 
         # TODO: we support rna only at present.
         table = meta.dataframe.to_dict(orient = 'list')
@@ -81,7 +86,7 @@ class experiment:
 
                     self.modalities['rna'][i_sample] = read_table_rna(
                         src = i_loc, metadata = meta, sample = i_sample,
-                        raw = False, default_taxa = i_taxa
+                        raw = False, default_taxa = i_taxa, eccentric = eccentric
                     )
 
                 elif i_loc.endswith('.h5') or i_loc.endswith('.h5ad'):
@@ -94,7 +99,7 @@ class experiment:
                 else:
                     self.modalities['rna'][i_sample] = read_mtx_rna(
                         src = i_loc, prefix = '', metadata = meta, sample = i_sample,
-                        raw = False, default_taxa = i_taxa
+                        raw = False, default_taxa = i_taxa, eccentric = eccentric
                     )
                     
                 self.modalities['rna'][i_sample].var = \
@@ -158,33 +163,27 @@ class experiment:
             
             else: warning(f'sample {i_sample} have no supported modalities')
 
-        # construct variable name table: note that this variable table represent the
-        # all available set, and should be a superset of the merged table.
-        
-        self.build_variables()
-
         pass
 
 
-    def build_variables(self):
-
-        self.variables = {}
-
-        if len(self.modalities) == 0 and self.mudata is not None:
-            
-            if 'rna' in self.mudata.mod.keys():
-                self.variables['rna'] = self.mudata['rna'].var
-            
-            return
+    @property
+    def obs(self):
+        self.check_merged()
+        self.pull()
+        return self.mudata.obs
+    
 
-        if 'rna' in self.modalities.keys():
-            genes = []
-            for sample in self.modalities['rna'].keys():
-                genes = set(list(genes) + self.modalities['rna'][sample].var_names.tolist())
-            self.variables['rna'] = experiment.search_genes(genes)
-        
-        pass
+    @property
+    def var(self):
+        self.check_merged()
+        return self.mudata.var
+    
 
+    @property
+    def shape(self):
+        self.check_merged()
+        return (self.mudata.n_obs, self.mudata.n_vars)
+    
     
     @staticmethod
     def search_genes(genes):
@@ -508,6 +507,47 @@ class experiment:
         subset.write_h5mu(os.path.join(self.directory, 'subsets', subset_name + '.h5mu'))
 
 
+    def build_subsample(
+        self, subset_name, slot = 'rna',
+        key = None, n = 3000
+    ):
+        self.check_merged(slot)
+        self.pull()
+        from exprmat.dynamics.utils import indices_to_bool
+
+        if key is None:
+            if n > self.mudata.n_obs:
+                n = self.mudata.n_obs
+                warning('subsample n > obs. returning all indices')
+            else: info(f'randomly selects {n} cells from {self.mudata.n_obs}.')
+            rind = np.random.choice(self.mudata.n_obs, size = n, replace = False)
+            mask = indices_to_bool(rind, self.mudata.n_obs)
+
+        elif isinstance(n, dict):
+            values = self.mudata.obs[f'{slot}:{key}'].unique().tolist()
+            rind = np.array([])
+            for val in values:
+                indices = np.argwhere(self.mudata.obs[f'{slot}:{key}'] == val).flatten()
+                if str(val) in n.keys(): _n = n[str(val)]
+                else: continue
+                if _n == '*': _n = len(indices)
+
+                if _n > len(indices):
+                    _n = len(indices)
+                    warning(f'subsample n > len({val}). returning all indices')
+                else: info(f'randomly selects {_n} cells from [{val}] ({len(indices)}).')
+                rind = np.append(rind, np.random.choice(indices, size = _n, replace = False))
+            mask = indices_to_bool(rind, self.mudata.n_obs)
+
+        subset = self.mudata[mask, :].copy()
+        info(f'selected {mask.sum()} observations from {len(mask)}.')
+
+        if not os.path.exists(os.path.join(self.directory, 'subsets')):
+            os.makedirs(os.path.join(self.directory, 'subsets'))
+        
+        subset.write_h5mu(os.path.join(self.directory, 'subsets', subset_name + '.h5mu'))
+
+
     def annotate(
         self, slot = 'rna', annotation = 'cell.type',
         mapping = {}, cluster = 'leiden'
@@ -673,7 +713,7 @@ class experiment:
         ls = self.mudata[slot].obs[annotation].tolist()
         mask = [x not in remove for x in ls]
         orig = self.mudata.n_obs
-        self.mudata = self.mudata[mask, :]
+        self.mudata = self.mudata[mask, :].copy()
         info(f'keep {self.mudata.n_obs} observations from {orig}.')
 
 
@@ -1326,6 +1366,11 @@ class experiment:
         from exprmat.reduction.plot import gene_gene
         return gene_gene(adata, sample_name = sample_name, **kwargs)
     
+    @staticmethod
+    def rna_plot_gene_gene_regress(adata, sample_name, **kwargs):
+        from exprmat.reduction.plot import gene_gene_regress
+        return gene_gene_regress(adata, sample_name = sample_name, **kwargs)
+    
 
     @staticmethod
     def rna_plot_markers(adata, sample_name, figsize, dpi, **kwargs):
@@ -1735,26 +1780,29 @@ class experiment:
     @staticmethod
     def rna_plot_gsea_dotplot(
         adata, sample_name, gsea_key, max_fdr = 1, max_p = 0.05, top_term: int = 100,
-        colour = 'p', title = "", cmap = 'turbo', figsize = (3, 2), cutoff = 1, ptsize = 5
+        colour = 'p', title = "", cmap = 'turbo', figsize = (3, 2), cutoff = 1, ptsize = 5,
+        terms = None
     ):
         from exprmat.plotting.gse import gsea_dotplot
         return gsea_dotplot(
             experiment.rna_get_gsea(adata, gsea_key, max_fdr = max_fdr, max_p = max_p),
             column = colour, x = 'nes', y = 'name', title = gsea_key if title is None else title,
-            cmap = cmap, size = ptsize, figsize = figsize, cutoff = cutoff, top_term = top_term
+            cmap = cmap, size = ptsize, figsize = figsize, cutoff = cutoff, top_term = top_term,
+            terms = terms
         )
     
 
     @staticmethod
     def rna_plot_opa_dotplot(
-        adata, sample_name, opa_key, max_fdr = 1, max_p = 0.05, top_term: int = 100,
+        adata, sample_name, opa_key, max_fdr = 1, max_p = 0.05, top_term: int = 100, terms = None,
         colour = 'fdr', title = None, cmap = 'turbo', figsize = (3, 2), cutoff = 1, ptsize = 5
     ):
         from exprmat.plotting.gse import opa_dotplot
         return opa_dotplot(
             experiment.rna_get_opa(adata, opa_key, max_fdr = max_fdr, max_p = max_p),
             column = colour, x = 'or', y = 'term', title = opa_key if title is None else title,
-            cmap = cmap, size = ptsize, figsize = figsize, cutoff = cutoff, top_term = top_term
+            cmap = cmap, size = ptsize, figsize = figsize, cutoff = cutoff, top_term = top_term,
+            terms = terms
         )
     
 
@@ -1776,6 +1824,25 @@ class experiment:
         return heatmap(adata = adata, uns_key = lr_key, **kwargs)
     
 
+    @staticmethod
+    def rna_plot_volcano(
+        adata, sample_name, de_slot = 'deg', label = [],
+        show_all = False, min_pct = 0, max_pct_reference = 1, 
+        min_lfc = -25, max_lfc = 25, remove_zero_pval = False,
+        highlight_min_logp = 5, highlight_min_lfc = 1.5,
+        xlim = 5, ylim = 100,
+        figsize = (3, 3), dpi = 100, **kwargs):
+        from exprmat.plotting.expression import volcano
+        return volcano(
+            adata = adata, de_slot = de_slot, label = label,
+            show_all = show_all, min_pct = min_pct, max_pct_reference = max_pct_reference, 
+            min_lfc = min_lfc, max_lfc = max_lfc, remove_zero_pval = remove_zero_pval,
+            highlight_min_logp = highlight_min_logp, highlight_min_lfc = highlight_min_lfc,
+            xlim = xlim, ylim = ylim,
+            figsize = figsize, dpi = dpi, **kwargs
+        )
+
+
     @staticmethod
     def rna_plot_spliced_proportions(adata, sample_name, **kwargs):
         from exprmat.plotting.velocity import proportions
@@ -2107,6 +2174,9 @@ class experiment:
     
     def plot_rna_gene_gene(self, run_on_samples = False, **kwargs):
         return self.plot_for_rna(run_on_samples, experiment.rna_plot_gene_gene, **kwargs)
+    
+    def plot_rna_gene_gene_regress(self, run_on_samples = False, **kwargs):
+        return self.plot_for_rna(run_on_samples, experiment.rna_plot_gene_gene_regress, **kwargs)
 
     def plot_rna_gene_gene_multiple(self, run_on_samples = False, **kwargs):
         return self.plot_for_rna(run_on_samples, experiment.rna_plot_multiple_gene_gene, **kwargs)
@@ -2184,6 +2254,9 @@ class experiment:
     def plot_rna_lr_circleplot(self, run_on_samples = False, **kwargs):
         return self.plot_for_rna(run_on_samples, experiment.rna_plot_lr_circleplot, **kwargs)
     
+    def plot_rna_volcano(self, run_on_samples = False, **kwargs):
+        return self.plot_for_rna(run_on_samples, experiment.rna_plot_volcano, **kwargs)
+    
     def plot_rna_spliced_proportions(self, run_on_samples = False, **kwargs):
         return self.plot_for_rna(run_on_samples, experiment.rna_plot_spliced_proportions, **kwargs)
     
@@ -2531,5 +2604,4 @@ def load_experiment(direc, load_samples = True, load_subset = None):
         subset = subset
     )
 
-    expr.build_variables()
     return expr

{exprmat-0.1.15 → exprmat-0.1.16}/exprmat/reader/matcher.py

@@ -62,7 +62,7 @@ def refine_finder(features_fpath, taxa = 'mmu'):
         save_genome_changes(taxa)
 
 
-def adjust_features(path, refine_finder = False, default_taxa = 'mmu'):
+def adjust_features(path, refine_finder = False, default_taxa = 'mmu', eccentric = None):
     '''
     There are three main types of features table that may occur in the matrix data folder.
     
@@ -121,6 +121,7 @@ def adjust_features(path, refine_finder = False, default_taxa = 'mmu'):
             # and indices in multi-species joint analysis. if there is no such prefix detected
             # this means the gene comes from the default taxa.
             
+            if eccentric is not None: ens = eccentric(ens)
             if '_' in ens:
                 reference_name = ens.split('_')[0]
                 pure_ens = ens.replace(reference_name + '_', '')
@@ -164,7 +165,8 @@ def adjust_features(path, refine_finder = False, default_taxa = 'mmu'):
         query_ensembl = []
         n_not_found = 0
         for x in query_gname:
-
+            
+            if eccentric is not None: x = eccentric(x)
             if '_' in x:
                 
                 reference_name = x.split('_')[0]
@@ -219,10 +221,10 @@ def adjust_features(path, refine_finder = False, default_taxa = 'mmu'):
     
 
 def read_mtx_rna(
-        src: str, prefix: str, 
-        metadata: metadata, sample: str, raw: bool = False,
-        default_taxa = 'mmu'
-    ):
+    src: str, prefix: str, 
+    metadata: metadata, sample: str, raw: bool = False,
+    default_taxa = 'mmu', eccentric = None
+):
     '''
     Formalize the three components in the matrix directory, and read the folder with gene
     expression data only.
@@ -251,7 +253,7 @@ def read_mtx_rna(
 
     # try:
     
-    adjust_features(src, refine_finder = True, default_taxa = default_taxa)
+    adjust_features(src, refine_finder = True, default_taxa = default_taxa, eccentric = eccentric)
 
     import warnings
     warnings.filterwarnings('ignore')
@@ -261,6 +263,9 @@ def read_mtx_rna(
         prefix = prefix
     )
 
+    if eccentric is not None:
+        adata.var_names = [eccentric(x) for x in adata.var_names]
+
     # except Exception as ex: 
     #     warning('error occurs when reading matrix files:')
     #     warning(str(ex))
@@ -278,10 +283,10 @@ def read_mtx_rna(
 
 
 def read_table_rna(
-        src: str, 
-        metadata: metadata, sample: str, raw: bool = False,
-        default_taxa = 'mmu'
-    ):
+    src: str, 
+    metadata: metadata, sample: str, raw: bool = False,
+    default_taxa = 'mmu', eccentric = None
+):
     '''
     Read from compressed or plain text tables.
 
@@ -311,6 +316,9 @@ def read_table_rna(
         src, delimiter = '\t', first_column_names = True, dtype = 'float32'
     ).T
 
+    if eccentric is not None:
+        adata.var_names = [eccentric(x) for x in adata.var_names]
+
     # except Exception as ex: 
     #     warning('error occurs when reading matrix files:')
     #     warning(str(ex))