pid_over_time.py

import pickle
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from math import ceil

df_al = pickle.load(open("alignment.pickle", "rb"))
df_sum = pickle.load(open("summary.pickle", "rb"))

df_al_sel = df_al[df_al["lengths"] > 200][["readIDs", "percentIdentity"]].set_index('readIDs')
df_sum_sel = df_sum[["readIDs", "start_time"]].set_index('readIDs')
df = df_sum_sel.join(df_al_sel, how="inner")

maxtime = df["start_time"].max().total_seconds()

df['timebin'] = pd.cut(
    x=df["start_time"],
    bins=ceil((maxtime / 3600) / 3),
    labels=[str(i) + "-" + str(i + 3) for i in range(0, 168, 3) if not i > (maxtime / 3600)])
ax = sns.violinplot(
    x=df["timebin"],
    y=df["percentIdentity"],
    inner=None,
    cut=0,
    linewidth=0)
ax.set(
    xlabel='Interval (hours)',
    ylabel="Percent identity",
    title="Percent identity over time")
plt.xticks(rotation=45, ha='center', fontsize=8)
plt.ylim(50, 100)
fig = ax.get_figure()
fig.savefig(
    fname="percentIdentityOverTimeViolin.png",
    format="png",
    bbox_inches='tight')
plt.close("all")