FROM 4OYD

Question: Can we fold 3LHP and expect it to mimic the configuration of 4OYD?

In this scenario, we use 3LHP as template, and 4OYD as the source of our motif and binder and as the final target conformation we are aiming to mimic.

We will run the protocol with and without a binder. Using full CA constraints from the template (fullcst) or just constraints between residues belonging to the same secondary structure (ssecst).

We will first have a no_design run that entails only the NubInitioMover. This runs will generate poses that have to be considered as non-final, and, thus, should not be considered on their own without further processing.

After that, different design rules are tested:

classicFFL: Rules are followed like in the original FFL. 3 Rounds of FastDesign in which the binder (if any) and the motif are completely static.
minimizeBinderAlternate: For 3 rounds, we will alternate FastDesign with the classical FFL rules with minimization of the binder.

Nomenclature:

Template: Structure that will accomodate the query motif.
Design: Chain that we are designing; contains the query motif and is guided by the template info.
Binder: If any, is whatever protein(s) bind to the query motif.
Target: Protein that we expect our design to resemble to and to which we compare.



In [2]:

    
%load_ext autoreload
%autoreload 2



In [4]:

    
%matplotlib inline
import matplotlib
import matplotlib.pyplot as plt
import rstoolbox
import pandas as pd
import seaborn as sns
import numpy as np
import copy

sns.set(font_scale=1.5);



In [30]:

    
def merge_rmsd_types( df, keys ):
    dataframes = []
    for k in keys["split"]:
        colIDs = copy.copy(keys["keep"])
        colIDs.append(k[0])
        wdf = df[colIDs]
        wdf = wdf.assign(temporarykey1=pd.Series([k[1]]*len(wdf[colIDs[0]])).values).copy(True)
        wdf = wdf.rename(index=str, columns={
            k[0]: keys["names"][0],
            "temporarykey1": keys["names"][1]
        })
        if ( len(k) > 2 ):
            wdf = wdf.assign(temporarykey2=pd.Series([k[2]]*len(wdf[colIDs[0]])).values).copy(True)
            wdf = wdf.rename(index=str, columns={
                "temporarykey2": keys["names"][2]
            })
        dataframes.append(wdf)
    return pd.concat(dataframes)

classicFFL (ffl)



In [120]:

    
experiments = [
    ("classicFFL/binder/design_classicFFL_1_minisilent", "classicFFL/binder/ddg_evaluation.score"),
    ("classicFFL/nobinder/design_classicFFL_1_minisilent", "classicFFL/nobinder/ddg_evaluation.score")
]
definition = {
    "scores":{
        "description": "description", "design_score": "score",
        "GRMSD2Target": "GRMSD2Target",
        "LRMSD2Target": "LRMSD2Target", "LRMSDH2Target": "LRMSDH2Target", "LRMSDLH2Target": "LRMSDLH2Target"
    },
    "naming": ["", "", "binder", "", "cluster", "decoy", "experiment", "", "", ""]
}
ddg_definition = {
    "scores":{
        "description": "description", "ddg": "ddg", "post_ddg": "min_ddg",
        #"rmsd_drift": "rmsd_drift", "total_score": "dimer_score"
        "score": "dimer_score"
    },
    "naming": ["", "", "binder", "", "cluster", "decoy", "experiment", "", "", "", ""]
}
logic = {
    "keep": ["description", "binder", "cluster", "decoy", "experiment",
             #"ddg", "min_ddg", "score", "dimer_score","rmsd_drift"],
             "ddg", "min_ddg", "score", "dimer_score"],
    "split": [("GRMSD2Target",  "global", "target"),  ("LRMSD2Target",   "local", "target"),
              ("LRMSDH2Target", "local",  "htarget"), ("LRMSDLH2Target", "local", "lhtarget")],
    "names": ["rmsd", "rmsd_alignment", "rmsd_to"]
}



In [121]:

    
dataframes = []
for experiment in experiments:
    dlist_main = rstoolbox.api.read_rosetta_silent(experiment[0])
    df_main    = rstoolbox.api.process_from_definitions(dlist_main, definition)
    dlist_ddg  = rstoolbox.api.read_rosetta_silent(experiment[1])
    df_ddg     = rstoolbox.api.process_from_definitions(dlist_ddg, ddg_definition)
    df_ddg     = df_ddg.drop("description", axis=1)
    df         = pd.merge(df_main, df_ddg, how='left', on=['cluster', 'decoy', 'experiment', 'binder'])
    dataframes.append(df)
ffl_data      = pd.concat(dataframes)
ffl_plot_data = merge_rmsd_types( ffl_data, logic )



In [82]:

    
g = sns.FacetGrid(ffl_plot_data, col="rmsd_to", size=9, aspect=0.5)
(g.map(sns.boxplot, "binder", "rmsd", showfliers=False)
 .despine(left=True))
plt.subplots_adjust(top=0.9)
g.fig.suptitle('RMSD with the Target (4OYD)')
sns.plt.show()



In [129]:

    
g = sns.lmplot(x="GRMSD2Target", y="LRMSDLH2Target", data=ffl_data, col="binder", fit_reg=False, size=10)
plt.subplots_adjust(top=0.80)
g.axes[0,0].set_ylim(0,5)
g.axes[0,0].set_xlim(0,3)
g.axes[0,0].set_yticks(np.arange(0,5.5,0.5))
g.fig.suptitle('4YOD as MOTIF SOURCE: classicFFL (LocalAlpha)')
sns.plt.show()



In [80]:

    
g = sns.lmplot(x="min_ddg", y="LRMSD2Target", col="binder", hue="binder", data=ffl_data, fit_reg=False, size=10)
plt.subplots_adjust(top=0.85)
g.axes[0,0].set_ylim(0, 5)
g.axes[0,0].set_xlim(-65,10)
g.fig.suptitle('ddG vs. LocalRMSD')
sns.plt.show()



In [127]:

    
g = sns.lmplot(x="min_ddg", y="score", col="binder", hue="binder", data=ffl_data, fit_reg=False, size=10)
plt.subplots_adjust(top=0.85)
g.axes[0,0].set_xlim(-70,10)
g.fig.suptitle('ddG vs. score')
sns.plt.show()



In [124]:

    
f, axes = plt.subplots(1, 3, figsize=(20, 9))
sns.boxplot(x="binder", y="score", data=ffl_data, showfliers=False, ax=axes[0])
sns.boxplot(x="binder", y="min_ddg", data=ffl_data, showfliers=False, ax=axes[1])
sns.boxplot(x="binder", y="dimer_score", data=ffl_data, showfliers=False, ax=axes[2])
plt.suptitle('ddG vs. score')
#f.subplots_adjust(top=0.85)
#plt.fig.suptitle('ddG vs. score')
sns.plt.show()



In [135]:

    
d = ffl_data[(ffl_data["binder"] == "binder")]
g = sns.lmplot(x="GRMSD2Target", y="LRMSD2Target", data=ffl_data, fit_reg=False, size=10)
plt.subplots_adjust(top=0.95)
g.axes[0,0].set_ylim(0,4)
g.axes[0,0].set_xlim(0,4)
#g.axes[0,0].set_yticks(np.arange(0,5.5,0.5))
sele_line_p1 = [0, 2.351]
sele_line_p2 = [1.605, 2.351]
selection_line = [
    [sele_line_p1[0], sele_line_p2[0]],
    [sele_line_p1[1], sele_line_p2[1]]
]
plt.plot(selection_line[0], selection_line[1], linewidth=2, color="red")
sele_line_p1 = [1.605, 2.351]
sele_line_p2 = [1.605, 0]
selection_line = [
    [sele_line_p1[0], sele_line_p2[0]],
    [sele_line_p1[1], sele_line_p2[1]]
]
plt.plot(selection_line[0], selection_line[1], linewidth=2, color="red")
g.fig.suptitle('4YOD as MOTIF SOURCE: classicFFL')
sns.plt.show()



In [141]:

    
d = ffl_data[(ffl_data["binder"] == "binder")]
g = sns.lmplot(x="GRMSD2Target", y="LRMSDLH2Target", data=ffl_data, fit_reg=False, size=10)
plt.subplots_adjust(top=0.95)
g.axes[0,0].set_ylim(0,4)
g.axes[0,0].set_xlim(0,4)
#g.axes[0,0].set_yticks(np.arange(0,5.5,0.5))
sele_line_p1 = [0, 2.351]
sele_line_p2 = [1.605, 2.351]
selection_line = [
    [sele_line_p1[0], sele_line_p2[0]],
    [sele_line_p1[1], sele_line_p2[1]]
]
plt.plot(selection_line[0], selection_line[1], linewidth=2, color="red")
sele_line_p1 = [1.605, 2.351]
sele_line_p2 = [1.605, 0]
selection_line = [
    [sele_line_p1[0], sele_line_p2[0]],
    [sele_line_p1[1], sele_line_p2[1]]
]
plt.plot(selection_line[0], selection_line[1], linewidth=2, color="red")
g.fig.suptitle('4YOD as MOTIF SOURCE: classicFFL (LocalAlpha)')
sns.plt.show()

minimizeBinderAlternate (mba)



In [137]:

    
experiments = "minimizeBinderAlternate/output/design_minimizeBinderAlternate_1_minisilent"
definition = {
    "scores":{
        "description": "description", "design_score": "score",
        "GRMSD2Target": "GRMSD2Target",
        "LRMSD2Target": "LRMSD2Target", "LRMSDH2Target": "LRMSDH2Target", "LRMSDLH2Target": "LRMSDLH2Target"
    },
    "naming": ["", "", "", "", "cluster", "decoy", "experiment", "", "", ""]
}



In [138]:

    
dlist    = rstoolbox.api.read_rosetta_silent(experiments)
mba_data = rstoolbox.api.process_from_definitions(dlist, definition)
mba_data.loc[mba_data["experiment"] == "minimizeBinderAlternate", "experiment"] = "mba"



In [142]:

    
g = sns.lmplot(x="GRMSD2Target", y="LRMSDLH2Target", data=mba_data, fit_reg=False, size=10)
plt.subplots_adjust(top=0.95)
g.axes[0,0].set_ylim(0,4)
g.axes[0,0].set_xlim(0,4)
#g.axes[0,0].set_yticks(np.arange(0,5.5,0.5))
sele_line_p1 = [0, 2.351]
sele_line_p2 = [1.605, 2.351]
selection_line = [
    [sele_line_p1[0], sele_line_p2[0]],
    [sele_line_p1[1], sele_line_p2[1]]
]
plt.plot(selection_line[0], selection_line[1], linewidth=2, color="red")
sele_line_p1 = [1.605, 2.351]
sele_line_p2 = [1.605, 0]
selection_line = [
    [sele_line_p1[0], sele_line_p2[0]],
    [sele_line_p1[1], sele_line_p2[1]]
]
plt.plot(selection_line[0], selection_line[1], linewidth=2, color="red")
g.fig.suptitle('4YOD as MOTIF SOURCE: minimize binding alternate (LocalAlpha)')
sns.plt.show()



In [140]:

    
g = sns.lmplot(x="GRMSD2Target", y="LRMSD2Target", data=mba_data, fit_reg=False, size=10)
plt.subplots_adjust(top=0.95)
g.axes[0,0].set_ylim(0,4)
g.axes[0,0].set_xlim(0,4)
#g.axes[0,0].set_yticks(np.arange(0,5.5,0.5))
sele_line_p1 = [0, 2.351]
sele_line_p2 = [1.605, 2.351]
selection_line = [
    [sele_line_p1[0], sele_line_p2[0]],
    [sele_line_p1[1], sele_line_p2[1]]
]
plt.plot(selection_line[0], selection_line[1], linewidth=2, color="red")
sele_line_p1 = [1.605, 2.351]
sele_line_p2 = [1.605, 0]
selection_line = [
    [sele_line_p1[0], sele_line_p2[0]],
    [sele_line_p1[1], sele_line_p2[1]]
]
plt.plot(selection_line[0], selection_line[1], linewidth=2, color="red")
g.fig.suptitle('4YOD as MOTIF SOURCE: minimize binding alternate')
sns.plt.show()



In [ ]: