QMplot: A Python tool for creating high-quality manhattan and Q-Q plots from GWAS results.

qmplot is a handy, user-friendly tool and Python library that allows for quick and flexible of publication-ready manhattan and Q-Q plots directly from PLINK association results files or any data frame with columns containing chromosomal name, chromosomal position, P-value and optionally the SNP name(e.g. rsID in dbSNP).

This library is inspired by r-qqman, but it's much more convenient than r-qqman that the column of chromosomal name doesn't have to be numeric any more, which means you can keep the raw name of chromosomal and don't have to covert X, Y, MT, etc to be 23, 24, 25, etc.

Dependencies

qmplot supports Python 3.6+ and no longer supports Python 2.

Instatllation requires numpy, scipy, pandas and matplotlib.

Installation

qmplot is written by Python and release in PyPI. The latest stable release can be installed by running the following command:

pip install qmplot

Quick Start

We use a PLINK2.x association output data "gwas_plink_result.tsv" which is in tests/data directory, as the input for the plots below. Here is the format preview of "gwas_plink_result.tsv":

#CHROM POS ID REF ALT A1 TEST OBS_CT BETA SE T_STAT P

1 904165 1_904165 G A A ADD 282 -0.0908897 0.195476 -0.464967 0.642344

1 1563691 1_1563691 T G G ADD 271 0.447021 0.422194 1.0588 0.290715

1 1707740 1_1707740 T G G ADD 283 0.149911 0.161387 0.928888 0.353805

1 2284195 1_2284195 T C C ADD 275 -0.024704 0.13966 -0.176887 0.859739

1 2779043 1_2779043 T C T ADD 272 -0.111771 0.139929 -0.79877 0.425182

1 2944527 1_2944527 G A A ADD 276 -0.054472 0.166038 -0.32807 0.743129

1 3803755 1_3803755 T C T ADD 283 -0.0392713 0.128528 -0.305547 0.760193

1 4121584 1_4121584 A G G ADD 279 0.120902 0.127063 0.951511 0.342239

1 4170048 1_4170048 C T T ADD 280 0.250807 0.143423 1.74873 0.0815274

qmplot apply two ways to generate manhattan and Q-Q plots:

1. Commandline options

This is the simplest way to plot manhattan and QQ plots if you already have PLINK2.x association output. You can directly type qmplot --help and will find all the options below:

usage: qmplot [-h] -I INPUT -O OUTPREFIX [-T TITLE] [-P SIGN_PVALUE] [-M M_ID]
          [--open-gui]

qmplot: Creates high-quality manhattan and QQ plots from PLINK association
output (or any dataframe with chromosome, position, and p-value).

optional arguments:
  -h, --help            show this help message and exit
  -I INPUT, --input INPUT
                        Input file
  -O OUTPREFIX, --outprefix OUTPREFIX
                        The prefix of output file
  -T TITLE, --title TITLE
                        Title of figure
  -P SIGN_PVALUE, --sign-mark-pvalue SIGN_PVALUE
                        Genome wide significant p-value sites. [1e-6]
  -M M_ID, --top-sign-signal-mark-id M_ID
                        A string denoting the column name for which you want
                        to annotate the Top Significant SNPs. Default: "ID"(PLINK2.x)
  --display             Set to be GUI backend, which can show the figure.

The following command will give you the two png plots with 300 dpi resolution:

$ qmplot -I data/gwas_plink_result.tsv -T Test -M ID --dpi 300 -O test

The manhattan plot looks like:

The Q-Q plot looks like:

Note: You can only modify the plots throught qmplot commandline options which is a big limitation when you want to make more change.

2. Python library

This is the most flexible way. You can use qmplot as a library in you Python code and create the plots by your mind.

Manhattan plot with default parameters

The manhattanplot() function in qmplot takes a data frame with columns containing the chromosomal name/id, chromosomal position, P-value and optionally the name of SNP(e.g. rsID in dbSNP).

By default, manhattanplot() looks for column names corresponding to those outout by the plink2 association results, namely, "#CHROM", "POS", "P", and "ID", although different column names can be specificed by user. Calling manhattanplot() function with a data frame of GWAS results as the single argument draws a basic manhattan plot, defaulting to a darkblue and lightblue color scheme.

import pandas as pd
import matplotlib.pyplot as plt
from qmplot import manhattanplot

if __name__ == "__main__":

    df = pd.read_table("tests/data/gwas_plink_result.tsv", sep="\t")
    df = df.dropna(how="any", axis=0)  # clean data

    # generate manhattan plot and set an output file.
    ax = manhattanplot(data=df)
    plt.savefig("output_manhattan_plot.png")

Rotate the x-axis tick label by setting xticklabel_kws to avoid label overlap:

ax = manhattanplot(data=df,
                   # set vertical or any other degrees as you like.
                   xticklabel_kws={"rotation": "vertical"})

Or rotate the labels 45 degrees by setting xticklabel_kws={"rotation": 45}.

When run with default parameters, the manhattanplot() function draws horizontal lines drawn at $-log_{10}{(10^{-5})}$ for "suggestive" associations and $-log_{10}{(5×10^{-8})}$ for the "genome-wide significant" threshold. These can be move to different locations or turned off completely with the arguments suggestiveline and genomewideline, respectively.

ax = manhattanplot(data=df,
                   suggestiveline=None,  # Turn off suggestiveline
                   genomewideline=None,  # Turn off genomewideline
                   xticklabel_kws={"rotation": "vertical"},
                   is_show=True)  # display the plot in screen

The behavior of the manhattanplot function changes slightly when results from only a single chromosome is used. Here, instead of plotting alternating colors and chromosome ID on the x-axis, the SNP's position on the chromosome is plotted on the x-axis:

# plot only results of chromosome 8.
manhattanplot(data=df, CHR="chr8", xlabel="Chromosome 8")
plt.savefig("output_chr8_manhattan_plot.png")

manhattanplot() funcion has the ability to highlight SNPs with significant GWAS signal and annotate the Top SNP, which has the lowest P-value:

ax = manhattanplot(data=df,
                   sign_marker_p=1e-6,  # highline the significant SNP with ``sign_marker_color`` color.
                   is_annotate_topsnp=True,  # annotate the top SNP
                   xticklabel_kws={"rotation": "vertical"})

Additionally, highlighting SNPs of interest can be combined with limiting to a single chromosome to enable "zooming" into a particular region containing SNPs of interest.

An example for a better Manhattan plot

Futher graphical parameters can be passed to the manhattanplot() function to control things like plot title, point character, size, colors, etc. Here is the example:

import pandas as pd
from qmplot import manhattanplot

if __name__ == "__main__":

    # loading data from local file
    df = pd.read_table("tests/data/gwas_plink_result.tsv", sep="\t")
    df = df.dropna(how="any", axis=0)  # clean data

    # defined the plot style
    f, ax = plt.subplots(figsize=(12, 4), facecolor='w', edgecolor='k')
    xtick = set(['chr' + i for i in list(map(str, range(1, 10))) + ['11', '13', '15', '18', '21', 'X']])
    manhattanplot(data=data,
                  marker=".",
                  sign_marker_p=1e-6,  # Genome wide significant p-value
                  sign_marker_color="r",
                  snp="ID",

                  title="Test",
                  xtick_label_set=xtick,
                  xlabel="Chromosome",
                  ylabel=r"$-log_{10}{(P)}$",

                  sign_line_cols=["#D62728", "#2CA02C"],
                  hline_kws={"linestyle": "--", "lw": 1.3},

                  is_annotate_topsnp=True,
                  ld_block_size=500000,  # 500000 bp
                  text_kws={"fontsize": 12,  # The fontsize of annotate text
                            "arrowprops": dict(arrowstyle="-", color="k", alpha=0.6)},
                  ax=ax)
    plt.savefig("better.manhattan.png")

Find more details about the parameters by typing manhattanplot? in IPython console.

QQ plot with defualt parameters

The qqplot() function can be used to generate a Q-Q plot to visualize the distribution of association "P-value". The qqplot() function takes a vector of P-values as its the only required argument.

import pandas as pd
from qmplot import qqplot

if __name__ == "__main__":

    df = pd.read_table("tests/data/gwas_plink_result.tsv", sep="\t")
    df = df.dropna(how="any", axis=0)  # clean data
    ax = qqplot(data=df["P"])

A better QQ plot

Futher graphical parameters can be passed to qqplot() to control the plot title, axis labels, point characters, colors, points sizes, etc. Here is the example:

import pandas as pd
import matplotlib.pyplot as plt
from qmplot import qqplot

if __name__ == "__main__":
    df = pd.read_table("tests/data/gwas_plink_result.tsv", sep="\t")
    df = df.dropna(how="any", axis=0)  # clean data

    # Create a Q-Q plot
    f, ax = plt.subplots(figsize=(6, 6), facecolor="w", edgecolor="k")
    qqplot(data=data["P"],
           marker="o",
           title="Test",
           xlabel=r"Expected $-log_{10}{(P)}$",
           ylabel=r"Observed $-log_{10}{(P)}$",
           ax=ax)

    plt.savefig("test.QQ.png")

Find more details about the parameters by typing qqplot? in IPython console.