gwaslab 1.0.2

A collection of handy tools for GWAS

gwaslab

A collection of handy python scripts for GWAS.

Just want to make lif eaiser and save myself from repetitive work.

What you can do with gwaslab:

1. Side-by-side Manhattan and QQ plot
2. Manhattan plot
3. QQ plot
4. Calculate lamda GC
5. [Select top SNPs based on a given window size.]
6. Convert beta/se <-> OR/95%L_U/95%L_L
7. Select hapmap3 SNPs from sumstats
8. Convert Observed scale heritability to liability scale heritability
9. read ldsc log and extract numeric results directly into a pandas dataframe.
10. compare the effect size of select variants / or automatically detected lead variants from two sumstats.

Requirements:

1. Python>3 2. "scipy" 3. "numpy" 4. "pandas" 5. "matplotlib" 6. "seaborn"

Install:

pip install gwaslab

Current version: 0.0.6

Usage:

Input: pandas dataframe

Create Manhattan plot and QQ plot with just one line

import gwaslab as gl

## creat qqplot and manhattan plot with just one line
## pass a dataframe in, and specify the column name for chromosome, base pair position, and also the p values.
gl.mqqplot(sumstats,"CHR","POS","PVALUE")

## adjust the plot, select top snps and add annotation sutomatically.
gl.mqqplot(sumstats,"CHR","POS","PVALUE",cut=20,cutfactor=10,anno=True,verbose=True,save=True,title="gwaslab")

## all options
gl.mqqplot(insumstats,
          chrom,
          pos,
          p,
          scaled=False,
          cut=0,
          cutfactor=10,
          cut_line_color="#ebebeb",
          windowsizekb=500,
          anno=None,
          sig_level=5e-8,
          sig_line_color="grey",
          suggestive_sig_level=5e-6,
          title =None,
          mtitle=None,
          qtitle=None,
          figsize =(15,5),
          fontsize = 10,
          colors = ["#000042", "#7878BA"],
          verbose=True,
          repel_force=0.03,
          gc=True,
          save=None,
          saveargs={"dpi":300,"facecolor":"white"}
          )

Or you can plot it separately.

Manhattan plot

gl.mplot()

QQ plot

gl.qqplot()

Calculate genomic inflation factor

gc(insumstats{"PVALUE"},mode="p",level=0.5)
gc(insumstats["Z"],mode="z",level=0.5)
gc(insumstats["chi2"],mode="chi2",level=0.5)

Extract top snps given a sliding window size

gl.getsig(insumstats,id,chrom,pos,p)

gl.getsig(insumstats,id,chrom,pos,p,windowsizekb=500,verbose=True,sig_level=5e-8)

Ref: Zhou, Wei, and Global Biobank Meta-analysis Initiative. "Global Biobank Meta-analysis Initiative: Powering genetic discovery across human diseases." medRxiv (2021).

Converting observed scale heritability to liability scale heritability

gl.h2_obs_to_liab(h2_obs, P, K)

gl.h2_obs_to_liab(h2_obs, P, K, se_obs=None)

Ref: Equation 23 Lee, Sang Hong, et al. "Estimating missing heritability for disease from genome-wide association studies." The American Journal of Human Genetics 88.3 (2011): 294-305.

Read ldsc results in to pandas DataFrame

Directly read ldsc -h2 or -rg into pandas dataframe...

pathlist=["./test.results.log","./test2.results.log"]

ldsc_h2 = gl.read_ldsc(pathlist, mode="h2")
ldsc_rg = gl.read_ldsc(pathlist, mode="rg")

ldsc_h2
Filename	h2_obs	h2_se	Lambda_gc	Mean_chi2	Intercept	Intercept_se	Ratio	Ratio_se
test.results.log	42.9954	8.657	1.2899	1.3226	0.0098	0.0098	0.6538	0.0304
test2.results.log	NA	NA	1.2899	1.3226	0.0098	0.0098	Ratio < 0	NA

ldsc_rg
p1	p2	rg	se	z	p	h2_obs	h2_obs_se	h2_int	h2_int_se	gcov_int	gcov_int_se
./test.results.log	./test.results.log	0.2317	0.0897	2.5824	0.0098	0.3305	0.0571	0.9612	0.009	-0.0001	0.0062
./test.results.log	./test2.results.log	0.2317	0.0897	2.5824	0.0098	0.3305	0.0571	0.9612	0.009	-0.0001	0.0062

Compare effect sizes of selected variants from two sumstats

gl.compare_effect()

preformat your sumstats for a qc workflow

---

Log

• 0.0.5 - 0.0.6

• added compare_effect, read_ldsc

• 0.0.4

  • added mqqplot feature
  • fixed gtesig algorithm
  • recreated mplot and qqplot

Next

• beta to OR
• OR to beta

For more information: https://gwaslab.com/