# This file contains configurations for matplotlib plotting parameters import matplotlib.pyplot as plt import matplotlib as mpl from matplotlib import font_manager from matplotlib.backends.backend_pdf import PdfPages from cycler import cycler import numpy as np import os colors = cycler('color', ['#4285F4', # Blue '#DB4437', # Red '#F4B400', # Yellow '#0F9D58', # Green '#904EE8', '#6B72B9', '#E377C3', '#FF8800', ]) config = {'font.family' : 'Arial', 'font.size' : 7, 'image.resample' : False, 'figure.dpi' : 600, 'xtick.major.size' : 2, 'ytick.major.size' : 2, 'xtick.major.width' : 0.5, 'ytick.major.width' : 0.5, 'axes.linewidth' : 0.5, 'lines.linewidth' : 0.5, 'savefig.dpi' : 600, 'savefig.transparent' : False, 'savefig.bbox' : 'tight', 'axes.prop_cycle' : colors, 'scatter.edgecolors' : 'none', 'legend.frameon' : False, } font_manager.fontManager.addfont('src/utils/Arial.ttf') mpl.rcParams.update(config) mpl.rcParams['pdf.fonttype'] = 42 def figure(figsize=(8, 6), font_size = 7, **kwargs): """ Create a figure with the given figsize and kwargs. figsize: tuple of (width, height) in millimeters """ figsize = (figsize[0] / 25.4, figsize[1] / 25.4) mpl.rcParams['font.size'] = font_size return plt.figure(figsize=figsize, **kwargs) def subplots(nrows=1, ncols=1, figsize=(8, 6), font_size = 7, **kwargs): """ Create a figure with the given figsize and kwargs. figsize: tuple of (width, height) in millimeters """ figsize = (figsize[0] / 25.4, figsize[1] / 25.4) mpl.rcParams['font.size'] = font_size fig, ax = plt.subplots(nrows=nrows, ncols=ncols, figsize=figsize, **kwargs) # Remove the top and right spines if nrows == 1 and ncols == 1: ax.spines['right'].set_visible(False) ax.spines['top'].set_visible(False) elif nrows == 1: for col in range(ncols): ax[col].spines['right'].set_visible(False) ax[col].spines['top'].set_visible(False) elif ncols == 1: for row in range(nrows): ax[row].spines['right'].set_visible(False) ax[row].spines['top'].set_visible(False) else: for row in range(nrows): for col in range(ncols): ax[row, col].spines['right'].set_visible(False) ax[row, col].spines['top'].set_visible(False) return fig, ax def plot_histogram( data, bins, plot_title, x_name, y_name, filename, data_save_path ): plt.figure(figsize=(8, 6)) plt.hist(data, bins=bins, color='#4285F4', edgecolor='black', alpha=0.7) plt.title(plot_title) plt.xlabel(x_name) plt.ylabel(y_name) plt.tight_layout() plt.savefig(os.path.join(data_save_path, f"{filename}.png"), bbox_inches='tight', dpi=300) plt.close() def plot_heatmap( plot_mat, plot_title, cbar_label, color_map, x_name, y_name, filename, data_save_path ): n_rows, n_cols = plot_mat.shape plt.figure(figsize=(8, 6)) plt.imshow(plot_mat, cmap=color_map, aspect='auto', vmin=0, vmax=1) plt.colorbar(label=cbar_label) plt.xlabel(x_name) plt.ylabel(y_name) plt.title(plot_title) # Show ALL ticks plt.xticks(range(n_cols)) plt.yticks(range(n_rows)) # Add text annotations for i in range(n_rows): for j in range(n_cols): plt.text(j, i, f'{plot_mat[i,j]:.2f}', ha='center', va='center', color='white' if plot_mat[i,j] > 0.5 else 'black', fontsize=8) plt.tight_layout() plt.savefig(os.path.join(data_save_path, f"{filename}.png"), bbox_inches='tight', dpi=300) plt.close() def plot_chipseq_tracks( data_mat, cap_list, loci_list=None, n_caps_per_page=100, track_height=0.3, figsize_width=300, smooth_window=None, log_scale=False, show_percentiles=False, data_save_path=".", filename_prefix="chipseq_tracks", output_format="pdf" # "pdf" for multipage PDF, "png" for separate files ): n_loci, n_caps = data_mat.shape n_pages = (n_caps + n_caps_per_page - 1) // n_caps_per_page # Create loci positions (x-axis) loci_positions = np.arange(n_loci) # Setup output format if output_format == "pdf": pdf_filename = os.path.join(data_save_path, f"{filename_prefix}.pdf") pdf_pages = PdfPages(pdf_filename) print(f"Creating multipage PDF: {pdf_filename}") for page in range(n_pages): start_cap = page * n_caps_per_page end_cap = min(start_cap + n_caps_per_page, n_caps) current_n_caps = end_cap - start_cap # Create figure figsize_height = current_n_caps * track_height * 25.4 # Convert to mm fig, axes = subplots( nrows=current_n_caps, ncols=1, figsize=(figsize_width, figsize_height), sharex=True ) # Handle single subplot case if current_n_caps == 1: axes = [axes] for i, cap_idx in enumerate(range(start_cap, end_cap)): ax = axes[i] cap_name = cap_list[cap_idx] # Get signal for this cap across all loci signal = data_mat[:, cap_idx] # Apply smoothing if requested if smooth_window is not None: signal = np.convolve(signal, np.ones(smooth_window)/smooth_window, mode='same') # Apply log scale if requested if log_scale: signal = np.log1p(signal) # log(1+x) to handle zeros # Plot the track - handle positive and negative values separately positive_mask = signal >= 0 negative_mask = signal < 0 # Plot positive values in blue (above x-axis) if np.any(positive_mask): pos_signal = np.where(positive_mask, signal, 0) ax.fill_between(loci_positions, 0, pos_signal, alpha=0.6, color='#4285F4') ax.plot(loci_positions, pos_signal, linewidth=0.3, alpha=0.8, color='#4285F4') # Plot negative values in red (below x-axis) if np.any(negative_mask): neg_signal = np.where(negative_mask, signal, 0) ax.fill_between(loci_positions, 0, neg_signal, alpha=0.6, color='#DB4437') ax.plot(loci_positions, neg_signal, linewidth=0.3, alpha=0.8, color='#DB4437') # Add percentile lines if requested if show_percentiles: p25, p50, p75 = np.percentile(signal, [25, 50, 75]) ax.axhline(y=p25, color='gray', linestyle='--', alpha=0.5, linewidth=0.3) ax.axhline(y=p50, color='red', linestyle='-', alpha=0.7, linewidth=0.3) ax.axhline(y=p75, color='gray', linestyle='--', alpha=0.5, linewidth=0.3) # Formatting - larger font for TF labels ax.set_ylabel(cap_name, rotation=0, ha='right', va='center', fontsize=8, fontweight='bold') # Set y-axis limits to show both positive and negative values signal_min, signal_max = np.min(signal), np.max(signal) y_margin = 0.05 * max(abs(signal_min), abs(signal_max)) ax.set_ylim(signal_min - y_margin, signal_max + y_margin) # Add horizontal line at y=0 for reference ax.axhline(y=0, color='black', linewidth=0.5, alpha=0.7) # Remove y-axis ticks to save space ax.set_yticks([]) # Tighten spacing ax.margins(y=0.05) # # Add statistics text # stats_text = f'Mean: {np.mean(signal):.2f}\nMax: {np.max(signal):.2f}' # if show_percentiles: # stats_text += f'\nMedian: {p50:.2f}' # ax.text(0.98, 0.98, stats_text, transform=ax.transAxes, # verticalalignment='top', horizontalalignment='right', fontsize=6, # bbox=dict(boxstyle='round,pad=0.3', facecolor='white', alpha=0.8)) # Set x-axis label only on bottom subplot axes[-1].set_xlabel('Loci Index') axes[-1].set_xlim(0, n_loci-1) # Add title to top subplot title = f'ChIP-seq Tracks - Page {page+1}/{n_pages}' if log_scale: title += ' (Log Scale)' if smooth_window: title += f' (Smoothed, window={smooth_window})' axes[0].set_title(title) plt.tight_layout() # Save figure based on output format if output_format == "pdf": pdf_pages.savefig(fig, bbox_inches='tight', dpi=300) print(f"Added page {page+1}/{n_pages} to PDF") else: filename = f"{filename_prefix}_page{page+1:02d}.png" plt.savefig(os.path.join(data_save_path, filename), bbox_inches='tight', dpi=300) print(f"Saved {filename}") plt.close() # Close PDF if using PDF format if output_format == "pdf": pdf_pages.close() print(f"Multipage PDF saved: {pdf_filename}") print(f"Total pages: {n_pages}") print(f"Total TFs: {n_caps}") def plot_variability(variability_results: dict, save_path: str = None, show_labels: bool = True, max_labels: int = 50): """Plot variability scores for genes and CAPs Args: variability_results: Results from ChromnitronDataset.get_most_variable() save_path: Path to save the plot (optional) show_labels: Whether to show feature names on axes max_labels: Maximum number of labels to show before switching to rank display """ gene_scores = variability_results['gene_scores'] gene_names = variability_results['gene_names'] cap_scores = variability_results['cap_scores'] cap_names = variability_results['cap_names'] # Reverse arrays to show highest variability at the top gene_scores = gene_scores[::-1] gene_names = gene_names[::-1] cap_scores = cap_scores[::-1] cap_names = cap_names[::-1] # Get method info from results condition_used = variability_results['based_on'] # Reasonable figure sizing - cap the maximum height max_height = 16 # Maximum height in inches gene_height = min(max_height, max(8, len(gene_scores) * 0.02)) # Much smaller scaling cap_height = min(max_height, max(8, len(cap_scores) * 0.02)) fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(16, max(gene_height, cap_height))) # Plot gene variability bars1 = ax1.barh(range(len(gene_scores)), gene_scores, color='#4285F4') # Blue ax1.set_xlabel(f'Variability Score (Variance, based on {condition_used})') ax1.set_title(f'Top {len(gene_scores)} Most Variable Genes') ax1.invert_yaxis() # Conditionally show gene labels if show_labels and len(gene_names) <= max_labels: ax1.set_yticks(range(len(gene_names))) ax1.set_yticklabels(gene_names, fontsize=max(6, min(10, 300/len(gene_names)))) else: ax1.set_yticks([0, len(gene_names)//4, len(gene_names)//2, 3*len(gene_names)//4, len(gene_names)-1]) ax1.set_yticklabels([f'Rank {i+1}' for i in [0, len(gene_names)//4, len(gene_names)//2, 3*len(gene_names)//4, len(gene_names)-1]]) ax1.set_ylabel('Gene Rank (by variability)') # Add corner text box with top/bottom gene names if len(gene_scores) > 6: top_genes_text = "Top 3 genes:\n" + "\n".join([f"{i+1}. {gene_names[i]}: {gene_scores[i]:.3f}" for i in range(3)]) bottom_genes_text = "Bottom 3 genes:\n" + "\n".join([f"{len(gene_names)+i-2}. {gene_names[i]}: {gene_scores[i]:.3f}" for i in [-3, -2, -1]]) full_text = f"{top_genes_text}\n\n{bottom_genes_text}" ax1.text(0.02, 0.98, full_text, transform=ax1.transAxes, fontsize=8, verticalalignment='top', horizontalalignment='left', bbox=dict(boxstyle='round,pad=0.5', facecolor='lightblue', alpha=0.8)) # Plot CAP variability bars2 = ax2.barh(range(len(cap_scores)), cap_scores, color='#DB4437') # Red ax2.set_xlabel(f'Variability Score (Variance, based on {condition_used})') ax2.set_title(f'Top {len(cap_scores)} Most Variable CAPs') ax2.invert_yaxis() # Conditionally show CAP labels if show_labels and len(cap_names) <= max_labels: ax2.set_yticks(range(len(cap_names))) ax2.set_yticklabels(cap_names, fontsize=max(6, min(10, 300/len(cap_names)))) else: ax2.set_yticks([0, len(cap_names)//4, len(cap_names)//2, 3*len(cap_names)//4, len(cap_names)-1]) ax2.set_yticklabels([f'Rank {i+1}' for i in [0, len(cap_names)//4, len(cap_names)//2, 3*len(cap_names)//4, len(cap_names)-1]]) ax2.set_ylabel('CAP Rank (by variability)') # Add corner text box with top/bottom CAP names if len(cap_scores) > 6: top_caps_text = "Top 3 CAPs:\n" + "\n".join([f"{i+1}. {cap_names[i]}: {cap_scores[i]:.3f}" for i in range(3)]) bottom_caps_text = "Bottom 3 CAPs:\n" + "\n".join([f"{len(cap_names)+i-2}. {cap_names[i]}: {cap_scores[i]:.3f}" for i in [-3, -2, -1]]) full_text = f"{top_caps_text}\n\n{bottom_caps_text}" ax2.text(0.02, 0.98, full_text, transform=ax2.transAxes, fontsize=8, verticalalignment='top', horizontalalignment='left', bbox=dict(boxstyle='round,pad=0.5', facecolor='lightgreen', alpha=0.8)) plt.tight_layout() if save_path: # Create directory if it doesn't exist os.makedirs(os.path.dirname(save_path), exist_ok=True) plt.savefig(save_path, dpi=300, bbox_inches='tight') plt.close() else: plt.show()