PrivacyScanData/target/visualize/visualize_5_6.py

def ma(df="dataframe", lfc=None, ct_count=None, st_count=None, basemean=None, pv=None, lfc_thr=(1, 1), pv_thr=0.05,
       valpha=1, dotsize=8,markerdot="o", dim=(6, 5), r=300, show=False, color=("green", "grey", "red"), ar=0,
       figtype='png',axtickfontsize=9, axtickfontname="Arial", axlabelfontsize=9, axlabelfontname="Arial",
       axxlabel=None, axylabel=None, xlm=None, ylm=None, fclines=False, fclinescolor='#2660a4', legendpos='best',
       figname='ma', legendanchor=None, legendlabels=['significant up', 'not significant', 'significant down'],
       plotlegend=False, theme=None, geneid=None, genenames=None, gfont=8, gstyle=1, title=None):
    _x, _y = 'A', 'M'
    assert General.check_for_nonnumeric(df[lfc]) == 0, 'dataframe contains non-numeric values in lfc column'
    if ct_count and st_count:
        assert General.check_for_nonnumeric(df[ct_count]) == 0, \
            'dataframe contains non-numeric values in ct_count column'
        assert General.check_for_nonnumeric(
            df[st_count]) == 0, 'dataframe contains non-numeric values in ct_count column'
    if basemean:
        assert General.check_for_nonnumeric(df[basemean]) == 0, \
            'dataframe contains non-numeric values in basemean column'

    # this is important to check if color or A exists and drop them as if you run multiple times same command
    # it may update old instance of df
    df = df.drop(['color_add_axy', 'A_add_axy'], axis=1, errors='ignore')
    assert len(set(color)) == 3, 'unique color must be size of 3'
    df.loc[(df[lfc] >= lfc_thr[0]) & (df[pv] < pv_thr), 'color_add_axy'] = color[0]  # upregulated
    df.loc[(df[lfc] <= -lfc_thr[1]) & (df[pv] < pv_thr), 'color_add_axy'] = color[2]  # downregulated
    df['color_add_axy'].fillna(color[1], inplace=True)  # intermediate
    if basemean:
        # basemean (mean of normalized counts from DESeq2 results)
        df['A_add_axy'] = df[basemean]
    else:
        df['A_add_axy'] = (np.log2(df[ct_count]) + np.log2(df[st_count])) / 2
    # plot
    assign_values = {col: i for i, col in enumerate(color)}
    color_result_num = [assign_values[i] for i in df['color_add_axy']]
    assert len(
        set(color_result_num)) == 3, 'either significant or non-significant genes are missing; try to change lfc_thr' \
                                     ' to include both significant and non-significant genes'
    if theme:
        General.style_bg(theme)
    plt.subplots(figsize=dim)
    if plotlegend:
        s = plt.scatter(df['A_add_axy'], df[lfc], c=color_result_num, cmap=ListedColormap(color),
                    alpha=valpha, s=dotsize, marker=markerdot)
        assert len(legendlabels) == 3, 'legendlabels must be size of 3'
        plt.legend(handles=s.legend_elements()[0], labels=legendlabels, loc=legendpos,
                       bbox_to_anchor=legendanchor)
    else:
        plt.scatter(df['A_add_axy'], df[lfc], c=color_result_num, cmap=ListedColormap(color),
                    alpha=valpha, s=dotsize, marker=markerdot)
    # draw a central line at M=0
    plt.axhline(y=0, color='#7d7d7d', linestyle='--')
    # draw lfc threshold lines
    if fclines:
        plt.axhline(y=lfc_thr[0], color=fclinescolor, linestyle='--')
        plt.axhline(y=-lfc_thr[1], color=fclinescolor, linestyle='--')
    if axxlabel:
        _x = axxlabel
    if axylabel:
        _y = axylabel
    GeneExpression.geneplot_ma(df, geneid, lfc, lfc_thr, genenames, gfont, gstyle)
    General.axis_labels(_x, _y, axlabelfontsize, axlabelfontname)
    General.axis_ticks(xlm, ylm, axtickfontsize, axtickfontname, ar)
    General.get_figure(show, r, figtype, figname, theme, title)