Amino acid preferences in viral coiled-coil domains

In this notebook we'll use localpdb to analyze the amino acid preferences in the viral coiled-coil domains.

Imports¶

In [3]:

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt

from localpdb import PDB

import pandas as pd import seaborn as sns import matplotlib.pyplot as plt from localpdb import PDB

Initiate localpdb and load Socket data¶

In [5]:

lpdb = PDB('/ssd/db/localpdb', plugins=['Socket'])

lpdb = PDB('/ssd/db/localpdb', plugins=['Socket'])

Search for viral proteins with RCSB search API wrapper¶

In [6]:

lpdb.search(attribute='struct_keywords.pdbx_keywords', operator='contains_phrase', 
            value='VIRAL PROTEIN', no_hits=-1, select=True)

lpdb.search(attribute='struct_keywords.pdbx_keywords', operator='contains_phrase', value='VIRAL PROTEIN', no_hits=-1, select=True)

Select only entries than contain coiled-coil domain¶

In [7]:

lpdb.entries = lpdb.entries[lpdb.entries['socket_7.0'].notnull()]

lpdb.entries = lpdb.entries[lpdb.entries['socket_7.0'].notnull()]

Select X-ray entries of reasonable quality¶

In [8]:

lpdb.entries = lpdb.entries.query('method == "diffraction"')
lpdb.entries = lpdb.entries.query('resolution <= 2.0')

lpdb.entries = lpdb.entries.query('method == "diffraction"') lpdb.entries = lpdb.entries.query('resolution <= 2.0')

Get Socket data for selected entries¶

In [9]:

socket = lpdb.get_socket_dict(method='heptads')

socket = lpdb.get_socket_dict(method='heptads')

Iterate over entries and collect per-residue statistics¶

In [13]:

stats = {pos: {aa: 0 for aa in 'AILVCDEFGHKMNPQRSTWYX'} for pos in 'abcdefg'}

# Iterate over all coiled-coil domains and compute stats
for entry, ccs in socket.items():
    for cc in ccs.values():
        for seq, hept_register in zip(cc['sequences'].values(), cc['heptads'].values()):
            for aa, hept in zip(seq, hept_register):
                stats[hept][aa] += 1

stats = {pos: {aa: 0 for aa in 'AILVCDEFGHKMNPQRSTWYX'} for pos in 'abcdefg'} # Iterate over all coiled-coil domains and compute stats for entry, ccs in socket.items(): for cc in ccs.values(): for seq, hept_register in zip(cc['sequences'].values(), cc['heptads'].values()): for aa, hept in zip(seq, hept_register): stats[hept][aa] += 1

Plot residue preferences among heptad positions¶

In [23]:

df = pd.DataFrame.from_dict(stats)
plt.figure(figsize=(7, 7))
ax = sns.heatmap(df, cmap='YlGnBu', cbar=False, linewidths=.1)
ax.set_xticklabels(ax.get_xmajorticklabels(), fontsize = 18)
ax.set_yticklabels(ax.get_ymajorticklabels(), fontsize = 16, rotation=45);

df = pd.DataFrame.from_dict(stats) plt.figure(figsize=(7, 7)) ax = sns.heatmap(df, cmap='YlGnBu', cbar=False, linewidths=.1) ax.set_xticklabels(ax.get_xmajorticklabels(), fontsize = 18) ax.set_yticklabels(ax.get_ymajorticklabels(), fontsize = 16, rotation=45);