LMSstat_python: Python implementation of LMSstat (https://github.com/SNUFML/LMSstat)

Installation

pip install lmsstat

Usage

Input format: the first two columns must be Sample (col 0) and Group (col 1). The remaining columns are treated as metabolites/features.

t-test, u-test, ANOVA, and Kruskal-Wallis test

from lmsstat import stat
import pandas as pd

data = pd.read_csv("data.csv")
result = stat.allstats(data)
# result = stat.allstats(data, p_adj=False) # When you don't want to adjust p-value

result.to_csv('result.csv', index=False)  # Save the result as a csv file

Normality test

보정되지 않은 결과이므로 주의.

from lmsstat import stat
import pandas as pd

path = "data.csv"
data = pd.read_csv(path)

result = stat.norm_test(data)
result

Data Standardization

import pandas as pd
from lmsstat import stat

path = "data.csv"
data = pd.read_csv(path)
scaled_data = stat.scaling(data)
scaled_data.to_csv("scaled_data.csv")

scaled_data

PCA

from lmsstat import plot
import pandas as pd

data = pd.read_csv("data.csv")

pca_plt = plot.plot_pca(data, n_components=2)
pca_plt[0].show()
print(f"R2: {pca_plt[1]}, Q2: {pca_plt[2]}")  # R2, Q2

PLS-DA

from lmsstat import plot
import pandas as pd

data = pd.read_csv("data.csv")

plsda_plt = plot.plot_plsda(data, n_components=2)
plsda_plt[0].show()
print(f"R2X: {plsda_plt[1]}, R2Y: {plsda_plt[2]}, Q2: {plsda_plt[3]}")  # R2, Q2
print(plsda_plt[4].head())  # VIP ranking (DataFrame)

Box plot, Bar plot

각각 현재 작업 디렉토리 밑에 만들어진 boxplot, barplot 폴더에 자동으로 저장됨.

from lmsstat import plot, stat
import pandas as pd

data = pd.read_csv("data.csv")

stats_res = stat.allstats(data)

plot.plot_box(data, stats_res, test_type="t-test")
plot.plot_bar(data, stats_res, test_type="t-test")

# plot only significant metabolites for the selected test (default alpha=0.05)
# plot.plot_box(data, stats_res, test_type="t-test", significant_only=True)
# plot.plot_bar(data, stats_res, test_type="t-test", significant_only=True, alpha=0.01)

# limit parallelism (use 1 to run serially)
# plot.plot_box(data, stats_res, test_type="t-test", max_workers=1)
# plot.plot_bar(data, stats_res, test_type="t-test", max_workers=2)

# if your environment shows growing memory usage with long runs, you can restart workers periodically
# plot.plot_box(data, stats_res, test_type="t-test", max_workers=4, restart_every=500)

# boxplot point rendering options
# plot.plot_box(data, stats_res, test_type="t-test", points="none")
# plot.plot_box(data, stats_res, test_type="t-test", points="non_outliers")

Heatmap

from lmsstat import plot
import pandas as pd

data = pd.read_csv("data.csv")

plot.plot_heatmap(data)