GnnForPrivacyScan/data/purposeSplit/visualize/visualize_5_27.py

def mhat(df="dataframe", chr=None, pv=None, log_scale=True, color=None, dim=(6,4), r=300, ar=90, gwas_sign_line=False,
         gwasp=5E-08, dotsize=8, markeridcol=None, markernames=None, gfont=8, valpha=1, show=False, figtype='png',
         axxlabel=None, axylabel=None, axlabelfontsize=9, axlabelfontname="Arial", axtickfontsize=9,
         axtickfontname="Arial", ylm=None, gstyle=1, figname='manhattan', theme=None):

    _x, _y = 'Chromosomes', r'$ -log_{10}(P)$'
    rand_colors = ('#a7414a', '#282726', '#6a8a82', '#a37c27', '#563838', '#0584f2', '#f28a30', '#f05837',
                   '#6465a5', '#00743f', '#be9063', '#de8cf0', '#888c46', '#c0334d', '#270101', '#8d2f23',
                   '#ee6c81', '#65734b', '#14325c', '#704307', '#b5b3be', '#f67280', '#ffd082', '#ffd800',
                   '#ad62aa', '#21bf73', '#a0855b', '#5edfff', '#08ffc8', '#ca3e47', '#c9753d', '#6c5ce7',
                   '#a997df', '#513b56', '#590925', '#007fff', '#bf1363', '#f39237', '#0a3200', '#8c271e')
    if log_scale:
        # minus log10 of P-value
        df['tpval'] = -np.log10(df[pv])
    else:
        # for Fst values
        df['tpval'] = df[pv]
    # df = df.sort_values(chr)
    # if the column contains numeric strings
    df = df.loc[pd.to_numeric(df[chr], errors='coerce').sort_values().index]
    # add indices
    df['ind'] = range(len(df))
    df_group = df.groupby(chr)
    if color is not None and len(color) == 2:
        color_1 = int(df[chr].nunique() / 2) * [color[0]]
        color_2 = int(df[chr].nunique() / 2) * [color[1]]
        if df[chr].nunique() % 2 == 0:
            color_list = list(reduce(lambda x, y: x+y, zip(color_1, color_2)))
        elif df[chr].nunique() % 2 == 1:
            color_list = list(reduce(lambda x, y: x+y, zip(color_1, color_2)))
            color_list.append(color[0])
    elif color is not None and len(color) == df[chr].nunique():
        color_list = color
    elif color is None:
        # select colors randomly from the list based in number of chr
        color_list = sample(rand_colors, df[chr].nunique())
    else:
        print("Error: in color argument")
        sys.exit(1)

    xlabels = []
    xticks = []
    if theme == 'dark':
        general.dark_bg()
    fig, ax = plt.subplots(figsize=dim)
    i = 0
    for label, df1 in df.groupby(chr):
        df1.plot(kind='scatter', x='ind', y='tpval', color=color_list[i], s=dotsize, alpha=valpha, ax=ax)
        df1_max_ind = df1['ind'].iloc[-1]
        df1_min_ind = df1['ind'].iloc[0]
        xlabels.append(label)
        xticks.append((df1_max_ind - (df1_max_ind - df1_min_ind) / 2))
        i += 1

    # add GWAS significant line
    if gwas_sign_line is True:
        ax.axhline(y=-np.log10(gwasp), linestyle='--', color='#7d7d7d', linewidth=1)
    if markernames is not None:
        marker.geneplot_mhat(df, markeridcol, chr, pv, gwasp, markernames, gfont, gstyle, ax=ax)
    ax.margins(x=0)
    ax.margins(y=0)
    ax.set_xticks(xticks)
    if log_scale:
        ax.set_ylim([0, max(df['tpval'] + 1)])
    if ylm:
        ylm = np.arange(ylm[0], ylm[1], ylm[2])
    else:
        ylm = np.arange(0, max(df['tpval']+1), 1)
    ax.set_yticks(ylm)
    ax.set_xticklabels(xlabels, rotation=ar)
    # ax.set_yticklabels(ylm, fontsize=axtickfontsize, fontname=axtickfontname, rotation=ar)
    if axxlabel:
        _x = axxlabel
    if axylabel:
        _y = axylabel
    ax.set_xlabel(_x, fontsize=axlabelfontsize, fontname=axlabelfontname)
    ax.set_ylabel(_y, fontsize=axlabelfontsize, fontname=axlabelfontname)
    general.get_figure(show, r, figtype, figname, theme)