In [1]:

    

import os
import sys
import random
import time
from random import seed, randint
import argparse
import platform
from datetime import datetime
import imp
import numpy as np
import fileinput
from itertools import product
import pandas as pd
from scipy.interpolate import griddata
from scipy.interpolate import interp2d
import seaborn as sns
from os import listdir

import matplotlib.pyplot as plt
import seaborn as sns
from scipy.interpolate import griddata
import matplotlib as mpl
# sys.path.insert(0,'..')
# from notebookFunctions import *
# from .. import notebookFunctions

from Bio.PDB.PDBParser import PDBParser
from pyCodeLib import *


    

In [17]:

    

plt.rcParams['figure.figsize'] = np.array([16.18033, 10])    #golden ratio
plt.rcParams['figure.facecolor'] = 'w'
plt.rcParams['figure.dpi'] = 100
plt.rcParams.update({'font.size': 22})


    

In [5]:

    

fileLocation = "/Users/weilu/Research/simulation_stepups/dataset_membrane/native_energy/1iwg_membrane/info.dat"
data = pd.read_csv(fileLocation, sep="\s+")


    

In [19]:

    

dataset = {}
dataset["membrane"] = ["2bg9", "1j4n", "1py6_SD", "2bl2", "1rhz", "1iwg", "2ic8", "1pv6", "1occ", "1kpl", "2bs2", "1py6", "1u19"]
all_ = []
for pdb in dataset["membrane"]:
    fileLocation = f"/Users/weilu/Research/simulation_stepups/dataset_membrane/native_energy/{pdb}_membrane/info.dat"
    data = pd.read_csv(fileLocation, sep="\s+")
    all_.append(data.assign(Protein=pdb))
d_membrane = pd.concat(all_)
all_ = []
for pdb in dataset["membrane"]:
    fileLocation = f"/Users/weilu/Research/simulation_stepups/dataset_membrane/native_energy/{pdb}_globular/info.dat"
    data = pd.read_csv(fileLocation, sep="\s+")
    all_.append(data.assign(Protein=pdb))
d_globuar = pd.concat(all_)
d = pd.concat([d_membrane.assign(Frag="membrane"), d_globuar.assign(Frag="globular")]).reset_index(drop=True)


    

In [25]:

    

sns.scatterplot("Protein", "Fragment", hue="Frag", s=100, data=d)


    

    
Out[25]:





<matplotlib.axes._subplots.AxesSubplot at 0x1a238a8f28>






    

In [31]:

    

def getFromTerminal(CMD):
    return subprocess.Popen(CMD,stdout=subprocess.PIPE,shell=True).communicate()[0].decode()
def mySize(pdb):
    protein_length = getFromTerminal(f"wc /Users/weilu/Research/simulation_stepups/dataset_membrane/setups/{pdb}/ssweight").split()[0]
    return protein_length


    

In [34]:

    

d["Length"] = d["Protein"].apply(lambda x: mySize(x))


    

In [37]:

    

d["Fragment_per_aa"] = d["Fragment"]/d["Length"].astype(int)


    

In [39]:

    

sns.scatterplot("Protein", "Fragment_per_aa", hue="Frag", s=100, data=d)


    

    
Out[39]:





<matplotlib.axes._subplots.AxesSubplot at 0x1a1fcf4198>






    

In [4]:

    

pdb = "/Users/weilu/Research/examples/openMM_compare_energy/model_062_noDNA"


    

In [5]:

    

structure = parse_pdb(pdb)


    

In [8]:

    

chains = list(structure.get_chains())


    

In [10]:

    

chain = chains[0]


    

In [16]:

    

res_list = list(chain.get_residues())
res_list = [residue for residue in res_list if not is_hetero(residue)]


    

In [18]:

    

res_list


    

    
Out[18]:





[<Residue MET het=  resseq=1 icode= >,
 <Residue GLU het=  resseq=2 icode= >,
 <Residue GLU het=  resseq=3 icode= >,
 <Residue PRO het=  resseq=4 icode= >,
 <Residue GLN het=  resseq=5 icode= >,
 <Residue SER het=  resseq=6 icode= >,
 <Residue ASP het=  resseq=7 icode= >,
 <Residue PRO het=  resseq=8 icode= >,
 <Residue SER het=  resseq=9 icode= >,
 <Residue VAL het=  resseq=10 icode= >,
 <Residue GLU het=  resseq=11 icode= >,
 <Residue PRO het=  resseq=12 icode= >,
 <Residue PRO het=  resseq=13 icode= >,
 <Residue LEU het=  resseq=14 icode= >,
 <Residue SER het=  resseq=15 icode= >,
 <Residue GLN het=  resseq=16 icode= >,
 <Residue GLU het=  resseq=17 icode= >,
 <Residue THR het=  resseq=18 icode= >,
 <Residue PHE het=  resseq=19 icode= >,
 <Residue SER het=  resseq=20 icode= >,
 <Residue ASP het=  resseq=21 icode= >,
 <Residue LEU het=  resseq=22 icode= >,
 <Residue TRP het=  resseq=23 icode= >,
 <Residue LYS het=  resseq=24 icode= >,
 <Residue LEU het=  resseq=25 icode= >,
 <Residue LEU het=  resseq=26 icode= >,
 <Residue PRO het=  resseq=27 icode= >,
 <Residue GLU het=  resseq=28 icode= >,
 <Residue ASN het=  resseq=29 icode= >,
 <Residue ASN het=  resseq=30 icode= >,
 <Residue VAL het=  resseq=31 icode= >,
 <Residue LEU het=  resseq=32 icode= >,
 <Residue SER het=  resseq=33 icode= >,
 <Residue PRO het=  resseq=34 icode= >,
 <Residue LEU het=  resseq=35 icode= >,
 <Residue PRO het=  resseq=36 icode= >,
 <Residue SER het=  resseq=37 icode= >,
 <Residue GLN het=  resseq=38 icode= >,
 <Residue ALA het=  resseq=39 icode= >,
 <Residue MET het=  resseq=40 icode= >,
 <Residue ASP het=  resseq=41 icode= >,
 <Residue ASP het=  resseq=42 icode= >,
 <Residue LEU het=  resseq=43 icode= >,
 <Residue MET het=  resseq=44 icode= >,
 <Residue LEU het=  resseq=45 icode= >,
 <Residue SER het=  resseq=46 icode= >,
 <Residue PRO het=  resseq=47 icode= >,
 <Residue ASP het=  resseq=48 icode= >,
 <Residue ASP het=  resseq=49 icode= >,
 <Residue ILE het=  resseq=50 icode= >,
 <Residue GLU het=  resseq=51 icode= >,
 <Residue GLN het=  resseq=52 icode= >,
 <Residue TRP het=  resseq=53 icode= >,
 <Residue PHE het=  resseq=54 icode= >,
 <Residue THR het=  resseq=55 icode= >,
 <Residue GLU het=  resseq=56 icode= >,
 <Residue ASP het=  resseq=57 icode= >,
 <Residue PRO het=  resseq=58 icode= >,
 <Residue GLY het=  resseq=59 icode= >,
 <Residue PRO het=  resseq=60 icode= >,
 <Residue ASP het=  resseq=61 icode= >,
 <Residue GLU het=  resseq=62 icode= >,
 <Residue ALA het=  resseq=63 icode= >,
 <Residue PRO het=  resseq=64 icode= >,
 <Residue ARG het=  resseq=65 icode= >,
 <Residue MET het=  resseq=66 icode= >,
 <Residue PRO het=  resseq=67 icode= >,
 <Residue GLU het=  resseq=68 icode= >,
 <Residue ALA het=  resseq=69 icode= >,
 <Residue ALA het=  resseq=70 icode= >,
 <Residue PRO het=  resseq=71 icode= >,
 <Residue PRO het=  resseq=72 icode= >,
 <Residue VAL het=  resseq=73 icode= >,
 <Residue ALA het=  resseq=74 icode= >,
 <Residue PRO het=  resseq=75 icode= >,
 <Residue ALA het=  resseq=76 icode= >,
 <Residue PRO het=  resseq=77 icode= >,
 <Residue ALA het=  resseq=78 icode= >,
 <Residue ALA het=  resseq=79 icode= >,
 <Residue PRO het=  resseq=80 icode= >,
 <Residue THR het=  resseq=81 icode= >,
 <Residue PRO het=  resseq=82 icode= >,
 <Residue ALA het=  resseq=83 icode= >,
 <Residue ALA het=  resseq=84 icode= >,
 <Residue PRO het=  resseq=85 icode= >,
 <Residue ALA het=  resseq=86 icode= >,
 <Residue PRO het=  resseq=87 icode= >,
 <Residue ALA het=  resseq=88 icode= >,
 <Residue PRO het=  resseq=89 icode= >,
 <Residue SER het=  resseq=90 icode= >,
 <Residue TRP het=  resseq=91 icode= >,
 <Residue PRO het=  resseq=92 icode= >,
 <Residue LEU het=  resseq=93 icode= >,
 <Residue SER het=  resseq=94 icode= >,
 <Residue SER het=  resseq=95 icode= >,
 <Residue SER het=  resseq=96 icode= >,
 <Residue VAL het=  resseq=97 icode= >,
 <Residue PRO het=  resseq=98 icode= >,
 <Residue SER het=  resseq=99 icode= >,
 <Residue GLN het=  resseq=100 icode= >,
 <Residue LYS het=  resseq=101 icode= >,
 <Residue THR het=  resseq=102 icode= >,
 <Residue TYR het=  resseq=103 icode= >,
 <Residue GLN het=  resseq=104 icode= >,
 <Residue GLY het=  resseq=105 icode= >,
 <Residue SER het=  resseq=106 icode= >,
 <Residue TYR het=  resseq=107 icode= >,
 <Residue GLY het=  resseq=108 icode= >,
 <Residue PHE het=  resseq=109 icode= >,
 <Residue ARG het=  resseq=110 icode= >,
 <Residue LEU het=  resseq=111 icode= >,
 <Residue GLY het=  resseq=112 icode= >,
 <Residue PHE het=  resseq=113 icode= >,
 <Residue LEU het=  resseq=114 icode= >,
 <Residue HIS het=  resseq=115 icode= >,
 <Residue SER het=  resseq=116 icode= >,
 <Residue GLY het=  resseq=117 icode= >,
 <Residue THR het=  resseq=118 icode= >,
 <Residue ALA het=  resseq=119 icode= >,
 <Residue LYS het=  resseq=120 icode= >,
 <Residue SER het=  resseq=121 icode= >,
 <Residue VAL het=  resseq=122 icode= >,
 <Residue THR het=  resseq=123 icode= >,
 <Residue CYS het=  resseq=124 icode= >,
 <Residue THR het=  resseq=125 icode= >,
 <Residue TYR het=  resseq=126 icode= >,
 <Residue SER het=  resseq=127 icode= >,
 <Residue PRO het=  resseq=128 icode= >,
 <Residue ALA het=  resseq=129 icode= >,
 <Residue LEU het=  resseq=130 icode= >,
 <Residue ASN het=  resseq=131 icode= >,
 <Residue LYS het=  resseq=132 icode= >,
 <Residue MET het=  resseq=133 icode= >,
 <Residue PHE het=  resseq=134 icode= >,
 <Residue CYS het=  resseq=135 icode= >,
 <Residue GLN het=  resseq=136 icode= >,
 <Residue LEU het=  resseq=137 icode= >,
 <Residue ALA het=  resseq=138 icode= >,
 <Residue LYS het=  resseq=139 icode= >,
 <Residue THR het=  resseq=140 icode= >,
 <Residue CYS het=  resseq=141 icode= >,
 <Residue PRO het=  resseq=142 icode= >,
 <Residue VAL het=  resseq=143 icode= >,
 <Residue GLN het=  resseq=144 icode= >,
 <Residue LEU het=  resseq=145 icode= >,
 <Residue TRP het=  resseq=146 icode= >,
 <Residue VAL het=  resseq=147 icode= >,
 <Residue ASP het=  resseq=148 icode= >,
 <Residue SER het=  resseq=149 icode= >,
 <Residue THR het=  resseq=150 icode= >,
 <Residue PRO het=  resseq=151 icode= >,
 <Residue PRO het=  resseq=152 icode= >,
 <Residue PRO het=  resseq=153 icode= >,
 <Residue GLY het=  resseq=154 icode= >,
 <Residue THR het=  resseq=155 icode= >,
 <Residue ARG het=  resseq=156 icode= >,
 <Residue VAL het=  resseq=157 icode= >,
 <Residue ARG het=  resseq=158 icode= >,
 <Residue ALA het=  resseq=159 icode= >,
 <Residue MET het=  resseq=160 icode= >,
 <Residue ALA het=  resseq=161 icode= >,
 <Residue ILE het=  resseq=162 icode= >,
 <Residue TYR het=  resseq=163 icode= >,
 <Residue LYS het=  resseq=164 icode= >,
 <Residue GLN het=  resseq=165 icode= >,
 <Residue SER het=  resseq=166 icode= >,
 <Residue GLN het=  resseq=167 icode= >,
 <Residue HIS het=  resseq=168 icode= >,
 <Residue MET het=  resseq=169 icode= >,
 <Residue THR het=  resseq=170 icode= >,
 <Residue GLU het=  resseq=171 icode= >,
 <Residue VAL het=  resseq=172 icode= >,
 <Residue VAL het=  resseq=173 icode= >,
 <Residue ARG het=  resseq=174 icode= >,
 <Residue ARG het=  resseq=175 icode= >,
 <Residue CYS het=  resseq=176 icode= >,
 <Residue PRO het=  resseq=177 icode= >,
 <Residue HIS het=  resseq=178 icode= >,
 <Residue HIS het=  resseq=179 icode= >,
 <Residue GLU het=  resseq=180 icode= >,
 <Residue ARG het=  resseq=181 icode= >,
 <Residue CYS het=  resseq=182 icode= >,
 <Residue SER het=  resseq=183 icode= >,
 <Residue ASP het=  resseq=184 icode= >,
 <Residue SER het=  resseq=185 icode= >,
 <Residue ASP het=  resseq=186 icode= >,
 <Residue GLY het=  resseq=187 icode= >,
 <Residue LEU het=  resseq=188 icode= >,
 <Residue ALA het=  resseq=189 icode= >,
 <Residue PRO het=  resseq=190 icode= >,
 <Residue PRO het=  resseq=191 icode= >,
 <Residue GLN het=  resseq=192 icode= >,
 <Residue HIS het=  resseq=193 icode= >,
 <Residue LEU het=  resseq=194 icode= >,
 <Residue ILE het=  resseq=195 icode= >,
 <Residue ARG het=  resseq=196 icode= >,
 <Residue VAL het=  resseq=197 icode= >,
 <Residue GLU het=  resseq=198 icode= >,
 <Residue GLY het=  resseq=199 icode= >,
 <Residue ASN het=  resseq=200 icode= >,
 <Residue LEU het=  resseq=201 icode= >,
 <Residue ARG het=  resseq=202 icode= >,
 <Residue VAL het=  resseq=203 icode= >,
 <Residue GLU het=  resseq=204 icode= >,
 <Residue TYR het=  resseq=205 icode= >,
 <Residue LEU het=  resseq=206 icode= >,
 <Residue ASP het=  resseq=207 icode= >,
 <Residue ASP het=  resseq=208 icode= >,
 <Residue ARG het=  resseq=209 icode= >,
 <Residue ASN het=  resseq=210 icode= >,
 <Residue THR het=  resseq=211 icode= >,
 <Residue PHE het=  resseq=212 icode= >,
 <Residue ARG het=  resseq=213 icode= >,
 <Residue HIS het=  resseq=214 icode= >,
 <Residue SER het=  resseq=215 icode= >,
 <Residue VAL het=  resseq=216 icode= >,
 <Residue VAL het=  resseq=217 icode= >,
 <Residue VAL het=  resseq=218 icode= >,
 <Residue PRO het=  resseq=219 icode= >,
 <Residue TYR het=  resseq=220 icode= >,
 <Residue GLU het=  resseq=221 icode= >,
 <Residue PRO het=  resseq=222 icode= >,
 <Residue PRO het=  resseq=223 icode= >,
 <Residue GLU het=  resseq=224 icode= >,
 <Residue VAL het=  resseq=225 icode= >,
 <Residue GLY het=  resseq=226 icode= >,
 <Residue SER het=  resseq=227 icode= >,
 <Residue ASP het=  resseq=228 icode= >,
 <Residue CYS het=  resseq=229 icode= >,
 <Residue THR het=  resseq=230 icode= >,
 <Residue THR het=  resseq=231 icode= >,
 <Residue ILE het=  resseq=232 icode= >,
 <Residue HIS het=  resseq=233 icode= >,
 <Residue TYR het=  resseq=234 icode= >,
 <Residue ASN het=  resseq=235 icode= >,
 <Residue TYR het=  resseq=236 icode= >,
 <Residue MET het=  resseq=237 icode= >,
 <Residue CYS het=  resseq=238 icode= >,
 <Residue ASN het=  resseq=239 icode= >,
 <Residue SER het=  resseq=240 icode= >,
 <Residue SER het=  resseq=241 icode= >,
 <Residue CYS het=  resseq=242 icode= >,
 <Residue MET het=  resseq=243 icode= >,
 <Residue GLY het=  resseq=244 icode= >,
 <Residue GLY het=  resseq=245 icode= >,
 <Residue MET het=  resseq=246 icode= >,
 <Residue ASN het=  resseq=247 icode= >,
 <Residue ARG het=  resseq=248 icode= >,
 <Residue ARG het=  resseq=249 icode= >,
 <Residue PRO het=  resseq=250 icode= >,
 <Residue ILE het=  resseq=251 icode= >,
 <Residue LEU het=  resseq=252 icode= >,
 <Residue THR het=  resseq=253 icode= >,
 <Residue ILE het=  resseq=254 icode= >,
 <Residue ILE het=  resseq=255 icode= >,
 <Residue THR het=  resseq=256 icode= >,
 <Residue LEU het=  resseq=257 icode= >,
 <Residue GLU het=  resseq=258 icode= >,
 <Residue ASP het=  resseq=259 icode= >,
 <Residue SER het=  resseq=260 icode= >,
 <Residue SER het=  resseq=261 icode= >,
 <Residue GLY het=  resseq=262 icode= >,
 <Residue ASN het=  resseq=263 icode= >,
 <Residue LEU het=  resseq=264 icode= >,
 <Residue LEU het=  resseq=265 icode= >,
 <Residue GLY het=  resseq=266 icode= >,
 <Residue ARG het=  resseq=267 icode= >,
 <Residue ASN het=  resseq=268 icode= >,
 <Residue SER het=  resseq=269 icode= >,
 <Residue PHE het=  resseq=270 icode= >,
 <Residue GLU het=  resseq=271 icode= >,
 <Residue VAL het=  resseq=272 icode= >,
 <Residue ARG het=  resseq=273 icode= >,
 <Residue VAL het=  resseq=274 icode= >,
 <Residue CYS het=  resseq=275 icode= >,
 <Residue ALA het=  resseq=276 icode= >,
 <Residue CYS het=  resseq=277 icode= >,
 <Residue PRO het=  resseq=278 icode= >,
 <Residue GLY het=  resseq=279 icode= >,
 <Residue ARG het=  resseq=280 icode= >,
 <Residue ASP het=  resseq=281 icode= >,
 <Residue ARG het=  resseq=282 icode= >,
 <Residue ARG het=  resseq=283 icode= >,
 <Residue THR het=  resseq=284 icode= >,
 <Residue GLU het=  resseq=285 icode= >,
 <Residue GLU het=  resseq=286 icode= >,
 <Residue GLU het=  resseq=287 icode= >,
 <Residue ASN het=  resseq=288 icode= >,
 <Residue LEU het=  resseq=289 icode= >,
 <Residue ARG het=  resseq=290 icode= >,
 <Residue LYS het=  resseq=291 icode= >,
 <Residue LYS het=  resseq=292 icode= >,
 <Residue GLY het=  resseq=293 icode= >,
 <Residue GLU het=  resseq=294 icode= >,
 <Residue PRO het=  resseq=295 icode= >,
 <Residue HIS het=  resseq=296 icode= >,
 <Residue HIS het=  resseq=297 icode= >,
 <Residue GLU het=  resseq=298 icode= >,
 <Residue LEU het=  resseq=299 icode= >,
 <Residue PRO het=  resseq=300 icode= >,
 <Residue PRO het=  resseq=301 icode= >,
 <Residue GLY het=  resseq=302 icode= >,
 <Residue SER het=  resseq=303 icode= >,
 <Residue THR het=  resseq=304 icode= >,
 <Residue LYS het=  resseq=305 icode= >,
 <Residue ARG het=  resseq=306 icode= >,
 <Residue ALA het=  resseq=307 icode= >,
 <Residue LEU het=  resseq=308 icode= >,
 <Residue PRO het=  resseq=309 icode= >,
 <Residue ASN het=  resseq=310 icode= >,
 <Residue ASN het=  resseq=311 icode= >,
 <Residue THR het=  resseq=312 icode= >,
 <Residue SER het=  resseq=313 icode= >,
 <Residue SER het=  resseq=314 icode= >,
 <Residue SER het=  resseq=315 icode= >,
 <Residue PRO het=  resseq=316 icode= >,
 <Residue GLN het=  resseq=317 icode= >,
 <Residue PRO het=  resseq=318 icode= >,
 <Residue LYS het=  resseq=319 icode= >,
 <Residue LYS het=  resseq=320 icode= >,
 <Residue LYS het=  resseq=321 icode= >,
 <Residue PRO het=  resseq=322 icode= >,
 <Residue LEU het=  resseq=323 icode= >,
 <Residue ASP het=  resseq=324 icode= >,
 <Residue GLY het=  resseq=325 icode= >,
 <Residue GLU het=  resseq=326 icode= >,
 <Residue TYR het=  resseq=327 icode= >,
 <Residue PHE het=  resseq=328 icode= >,
 <Residue THR het=  resseq=329 icode= >,
 <Residue LEU het=  resseq=330 icode= >,
 <Residue GLN het=  resseq=331 icode= >,
 <Residue ILE het=  resseq=332 icode= >,
 <Residue ARG het=  resseq=333 icode= >,
 <Residue GLY het=  resseq=334 icode= >,
 <Residue ARG het=  resseq=335 icode= >,
 <Residue GLU het=  resseq=336 icode= >,
 <Residue ARG het=  resseq=337 icode= >,
 <Residue PHE het=  resseq=338 icode= >,
 <Residue GLU het=  resseq=339 icode= >,
 <Residue MET het=  resseq=340 icode= >,
 <Residue PHE het=  resseq=341 icode= >,
 <Residue ARG het=  resseq=342 icode= >,
 <Residue GLU het=  resseq=343 icode= >,
 <Residue LEU het=  resseq=344 icode= >,
 <Residue ASN het=  resseq=345 icode= >,
 <Residue GLU het=  resseq=346 icode= >,
 <Residue ALA het=  resseq=347 icode= >,
 <Residue LEU het=  resseq=348 icode= >,
 <Residue GLU het=  resseq=349 icode= >,
 <Residue LEU het=  resseq=350 icode= >,
 <Residue LYS het=  resseq=351 icode= >,
 <Residue ASP het=  resseq=352 icode= >,
 <Residue ALA het=  resseq=353 icode= >,
 <Residue GLN het=  resseq=354 icode= >,
 <Residue ALA het=  resseq=355 icode= >,
 <Residue GLY het=  resseq=356 icode= >,
 <Residue LYS het=  resseq=357 icode= >,
 <Residue GLU het=  resseq=358 icode= >,
 <Residue PRO het=  resseq=359 icode= >,
 <Residue GLY het=  resseq=360 icode= >,
 <Residue GLY het=  resseq=361 icode= >,
 <Residue SER het=  resseq=362 icode= >,
 <Residue ARG het=  resseq=363 icode= >,
 <Residue ALA het=  resseq=364 icode= >,
 <Residue HIS het=  resseq=365 icode= >,
 <Residue SER het=  resseq=366 icode= >,
 <Residue SER het=  resseq=367 icode= >,
 <Residue HIS het=  resseq=368 icode= >,
 <Residue LEU het=  resseq=369 icode= >,
 <Residue LYS het=  resseq=370 icode= >,
 <Residue SER het=  resseq=371 icode= >,
 <Residue LYS het=  resseq=372 icode= >,
 <Residue LYS het=  resseq=373 icode= >,
 <Residue GLY het=  resseq=374 icode= >,
 <Residue GLN het=  resseq=375 icode= >,
 <Residue SER het=  resseq=376 icode= >,
 <Residue THR het=  resseq=377 icode= >,
 <Residue SER het=  resseq=378 icode= >,
 <Residue ARG het=  resseq=379 icode= >,
 <Residue HIS het=  resseq=380 icode= >,
 <Residue LYS het=  resseq=381 icode= >,
 <Residue LYS het=  resseq=382 icode= >,
 <Residue LEU het=  resseq=383 icode= >,
 <Residue MET het=  resseq=384 icode= >,
 <Residue PHE het=  resseq=385 icode= >,
 <Residue LYS het=  resseq=386 icode= >,
 <Residue THR het=  resseq=387 icode= >,
 <Residue GLU het=  resseq=388 icode= >,
 <Residue GLY het=  resseq=389 icode= >,
 <Residue PRO het=  resseq=390 icode= >,
 <Residue ASP het=  resseq=391 icode= >,
 <Residue SER het=  resseq=392 icode= >,
 <Residue ASP het=  resseq=393 icode= >]

In [17]:

    

len(res_list)


    

    
Out[17]:





393

In [ ]:

    

for chain in chains:


    

In [19]:

    

pd.read_csv("/Users/weilu/Downloads/Table-S1-career-2017.xlsx")


    

    




---------------------------------------------------------------------------
ParserError                               Traceback (most recent call last)
<ipython-input-19-6978f19432a4> in <module>
----> 1 pd.read_csv("/Users/weilu/Downloads/Table-S1-career-2017.xlsx")

~/anaconda3/envs/py36/lib/python3.6/site-packages/pandas/io/parsers.py in parser_f(filepath_or_buffer, sep, delimiter, header, names, index_col, usecols, squeeze, prefix, mangle_dupe_cols, dtype, engine, converters, true_values, false_values, skipinitialspace, skiprows, skipfooter, nrows, na_values, keep_default_na, na_filter, verbose, skip_blank_lines, parse_dates, infer_datetime_format, keep_date_col, date_parser, dayfirst, iterator, chunksize, compression, thousands, decimal, lineterminator, quotechar, quoting, doublequote, escapechar, comment, encoding, dialect, tupleize_cols, error_bad_lines, warn_bad_lines, delim_whitespace, low_memory, memory_map, float_precision)
    700                     skip_blank_lines=skip_blank_lines)
    701 
--> 702         return _read(filepath_or_buffer, kwds)
    703 
    704     parser_f.__name__ = name

~/anaconda3/envs/py36/lib/python3.6/site-packages/pandas/io/parsers.py in _read(filepath_or_buffer, kwds)
    433 
    434     try:
--> 435         data = parser.read(nrows)
    436     finally:
    437         parser.close()

~/anaconda3/envs/py36/lib/python3.6/site-packages/pandas/io/parsers.py in read(self, nrows)
   1137     def read(self, nrows=None):
   1138         nrows = _validate_integer('nrows', nrows)
-> 1139         ret = self._engine.read(nrows)
   1140 
   1141         # May alter columns / col_dict

~/anaconda3/envs/py36/lib/python3.6/site-packages/pandas/io/parsers.py in read(self, nrows)
   1993     def read(self, nrows=None):
   1994         try:
-> 1995             data = self._reader.read(nrows)
   1996         except StopIteration:
   1997             if self._first_chunk:

pandas/_libs/parsers.pyx in pandas._libs.parsers.TextReader.read()

pandas/_libs/parsers.pyx in pandas._libs.parsers.TextReader._read_low_memory()

pandas/_libs/parsers.pyx in pandas._libs.parsers.TextReader._read_rows()

pandas/_libs/parsers.pyx in pandas._libs.parsers.TextReader._tokenize_rows()

pandas/_libs/parsers.pyx in pandas._libs.parsers.raise_parser_error()

ParserError: Error tokenizing data. C error: Expected 3 fields in line 51, saw 7

In [ ]: