def generate_2x1(consensus, rna=False):
seq = seqlib.standardize(consensus, rna)
listVariants = []
for i, base in enumerate(seq):
for otherBase1 in seqlib.allOtherBases(base, rna):
for otherBase2 in seqlib.allOtherBases(base, rna):
variant = seq[:i] + otherBase1 + otherBase2 + seq[i+1:]
listVariants.append(variant)
return listVariants
def generate_2x1(consensus, rna=False):
seq = seqlib.standardize(consensus, rna)
listVariants = []
for i, base in enumerate(seq):
for otherBase1 in seqlib.allOtherBases(base, rna):
for otherBase2 in seqlib.allOtherBases(base, rna):
variant = seq[:i] + otherBase1 + otherBase2 + seq[i + 1:]
listVariants.append(variant)
return listVariants
def generate_2x3(consensus, rna=False):
seq = seqlib.standardize(consensus, rna)
listVariants = []
for i, base in enumerate(seq[:-2]):
nextBase = seq[i + 1]
nextNextBase = seq[i + 2]
for otherBase1 in seqlib.allOtherBases(base, rna):
for otherBase3 in seqlib.allOtherBases(nextNextBase, rna):
variant = seq[:i] + otherBase1 + otherBase3 + seq[i + 3:]
listVariants.append(variant)
return listVariants
def generate_2x3(consensus, rna=False):
seq = seqlib.standardize(consensus, rna)
listVariants = []
for i, base in enumerate(seq[:-2]):
nextBase = seq[i+1]
nextNextBase = seq[i+2]
for otherBase1 in seqlib.allOtherBases(base, rna):
for otherBase3 in seqlib.allOtherBases(nextNextBase, rna):
variant = seq[:i] + otherBase1 + otherBase3 + seq[i+3:]
listVariants.append(variant)
return listVariants
def generate_4x4(consensus, rna=False):
seq = seqlib.standardize(consensus, rna)
listVariants = []
for i, base in enumerate(seq[:-3]):
nextBase = seq[i+1]
nextNextBase = seq[i+2]
nextNextNextBase = seq[i+3]
for otherBase1 in seqlib.allOtherBases(base, rna):
for newBase2 in seqlib.allBases(rna):
for newBase3 in seqlib.allBases(rna):
for otherBase4 in seqlib.allOtherBases(nextNextNextBase, rna):
variant = seq[:i] + otherBase1 + newBase2 + newBase3 + otherBase4 + seq[i+4:]
listVariants.append(variant)
return listVariants
def generate_4x4(consensus, rna=False):
seq = seqlib.standardize(consensus, rna)
listVariants = []
for i, base in enumerate(seq[:-3]):
nextBase = seq[i + 1]
nextNextBase = seq[i + 2]
nextNextNextBase = seq[i + 3]
for otherBase1 in seqlib.allOtherBases(base, rna):
for newBase2 in seqlib.allBases(rna):
for newBase3 in seqlib.allBases(rna):
for otherBase4 in seqlib.allOtherBases(
nextNextNextBase, rna):
variant = seq[:
i] + otherBase1 + newBase2 + newBase3 + otherBase4 + seq[
i + 4:]
listVariants.append(variant)
return listVariants
def generate_1x2(consensus, rna=False):
seq = seqlib.standardize(consensus, rna)
listVariants = []
for i, base in enumerate(seq[:-1]):
nextBase = seq[i + 1]
for otherBase in seqlib.allOtherBases([base, nextBase], rna):
variant = seq[:i] + otherBase + seq[i + 2:]
listVariants.append(variant)
return listVariants
def generate_1x2(consensus, rna=False):
seq = seqlib.standardize(consensus, rna)
listVariants = []
for i, base in enumerate(seq[:-1]):
nextBase = seq[i+1]
for otherBase in seqlib.allOtherBases([base, nextBase], rna):
variant = seq[:i] + otherBase + seq[i+2:]
listVariants.append(variant)
return listVariants
def doubleMutantMatrix(data, refVariant, libSeq, startPos=1, coop=False):
"""Auxiliary function to generate the doubel mutant matrix"""
# Get library positions and create labels for mutants
libPos = [i for (i, base) in enumerate(libSeq.upper()) if base == 'N']
mutantLabels = [refVariant[i]+str(i+startPos)+otherBase
for i in libPos for otherBase in seqlib.allOtherBases(refVariant[i])]
# Grep the mutants and fill in the signals
dim = len(mutantLabels)
doubleMutantSignals = np.zeros(shape=(dim, dim))
for i, mutation1 in enumerate(mutantLabels):
for j, mutation2 in enumerate(mutantLabels):
# Get the index for the degenerate base along the sequence
pos1 = int(mutation1[1:-1]) - startPos
otherBase1 = mutation1[-1]
pos2 = int(mutation2[1:-1]) - startPos
otherBase2 = mutation2[-1]
# Create the current mutant sequence
currSeq = list(refVariant)
currSeq[pos1] = otherBase1
currSeq[pos2] = otherBase2
currSeq = ''.join(currSeq)
# Grep the signal and fill in the double mutant matrix
if not (currSeq in data.index):
doubleMutantSignals[i, j] = np.nan
elif pos1 != pos2:
doubleMutantSignals[i, j] = data[currSeq]
elif (pos1 == pos2) and (otherBase1 == otherBase2):
doubleMutantSignals[i, j] = data[currSeq]
else:
doubleMutantSignals[i, j] = np.nan
# Compute cooperativity if requested
if coop:
coopSignals = np.zeros(shape=(dim, dim))
for i in xrange(dim):
for j in xrange(dim):
coopSignals[i, j] = doubleMutantSignals[i, i] + doubleMutantSignals[j, j] - doubleMutantSignals[i, j]
doubleMutantSignals = coopSignals
return doubleMutantSignals, mutantLabels
def plotTertSMcoop(dfUnqClusters, listConsensus, listSeqPos, refConsensus, field='params2.median', listConsensusName=None,
vmin=None, vmax=None, cmap='RdBu', c_bad='0.55', robust=True,
figsize=None, figunitheight=1, figunitwidth=1, maxfigwidth=32,
actLabel=r'$\mathrm{\mathsf{\Delta\Delta G^{\ddag}\ (kcal\ mol^{-1})}}$',
show=True, **kwargs):
"""Plot tertiary contact-single point mutant cooperivity heatmap"""
# Define constants
R = 1.9872041e-3
T = 293.0
# Make dfUnqClusters indexed by annotation if not already
if dfUnqClusters.index.name == 'annotation':
dfUnqClusters2 = dfUnqClusters.copy()
else:
dfUnqClusters2 = dfUnqClusters.set_index('annotation')
# Make an empty dataframe for the TC-SPM matrix
mat_kobs = pd.DataFrame(columns=listConsensus)
mat_ddG = pd.DataFrame(columns=listConsensus)
# Make an empty series for the consensus variants
seriesConsensusAct = pd.Series(index=listConsensus)
# Make the list of annotations to plot
listConsensusAct = []
for con in listConsensus:
listAnnt = []
listName = []
for pos in listSeqPos:
# Get base at the current position
currBase = pos[0]
# Get the 3 other bases at the current position
allOtherBases = seqlib.allOtherBases(currBase)
# Make the list of annotations of the 3 possible mismatches at the current position
mAnnts = [con+':1:0:0:'+pos+base+':::' for base in allOtherBases]
mNames = [pos+base for base in allOtherBases]
# Add to listAnnt and listName
listAnnt.extend(mAnnts)
listName.extend(mNames)
# Get the activity of the consensus variants
seriesConsensusAct[con] = 1./dfUnqClusters2.loc[con+':0:0:0::::'][field]*60
# Add the list of activity of the annotations from the current consensus as a column
mat_kobs[con] = (1./dfUnqClusters2.loc[listAnnt][field]*60).tolist()
# Compute delta delta G from k_obs
for con in listConsensus:
mat_ddG[con] = - R * T * np.log((mat_kobs[con] * seriesConsensusAct[refConsensus])
/ (mat_kobs[refConsensus] * seriesConsensusAct[con]))
#
if listConsensusName is not None:
mat_kobs.columns = listConsensusName
mat_ddG.columns = listConsensusName
#
mat_kobs.columns.name = 'Tertiary contact knockouts'
mat_ddG.columns.name = 'Tertiary contact knockouts'
# Set single mutant names
mat_kobs['Single point mutants'] = listName
mat_kobs = mat_kobs.set_index('Single point mutants').transpose()
mat_ddG['Single point mutants'] = listName
mat_ddG = mat_ddG.set_index('Single point mutants').transpose()
# Define colormap
if isinstance(cmap, basestring):
cmap = plt.get_cmap(cmap)
cmap.set_bad(c_bad, 0.8)
# Make mask for nan data
mask = ~np.isfinite(mat_ddG)
# Plot heatmap
if not figsize:
figwidth = min(figunitwidth * len(listAnnt), maxfigwidth)
figheight = figunitheight * len(listConsensus) + 0.5
figsize = (figwidth, figheight)
fig, ax = plt.subplots(1, 1, figsize=figsize)
cbar_ax = fig.add_axes([.78, .83, .2, .05])
sns.heatmap(mat_ddG, ax=ax, square=True, mask=mask, robust=robust,
vmin=vmin, vmax=vmax, center=0, cmap=cmap,
cbar_ax=cbar_ax, cbar_kws={'orientation': 'horizontal'})
setproperties(ax=cbar_ax, fontsize=21, xlabel=actLabel)
setproperties(ax=ax, tickfontsize=22, yticklabelrot=0,
labelfontsize=28, tight=True, pad=1.0)
if show:
plt.show(block=False)
return
def plotSingleMutants(df, consensus, listSeqPos, muttype='m',
colorbymut=True, fullname=False, collapse=False, **kwargs):
"""Plot the activities and counts of single mutants given a df of all clusters, name of consensus
sequence, and list of positions
This function defaults to plotting time/rate constants, but can be used to plot something else too"""
# Define constants
allBases = ['A', 'T', 'C', 'G']
allColors = {'A': sns.xkcd_rgb["windows blue"],
'T': sns.xkcd_rgb["amber"],
'C': sns.xkcd_rgb["faded green"],
'G': sns.xkcd_rgb["reddish orange"],
'x': sns.xkcd_rgb["dusty purple"]}
defaultColor = sns.xkcd_rgb["greyish"]
# Compute how many bars to be plotted at each position and xtickpos
numBarDict = {'m': 3, 'i': 3, 'd': 1}
numBar = np.sum([numBarDict.get(i, 0) for i in muttype])
gap = int(numBar) / 3
xtickspos = [range((numBar+gap)*pos+1, (numBar+gap)*pos+numBar+1) for pos in range(0, len(listSeqPos))]
xtickspos = np.array([item for sublist in xtickspos for item in sublist])
# Make the list of annotations to plot
listAnnt = []
listName = []
listColor = []
for pos in listSeqPos:
# Add mismatches
if 'm' in muttype:
# Get base at the current position
currBase = pos[0]
# Get the 3 other bases at the current position
allOtherBases = seqlib.allOtherBases(currBase)
# Make the list of annotations of the 3 possible mismatches at the current position
mAnnts = [consensus+':1:0:0:'+pos+base+':::' for base in allOtherBases]
mNames = [pos+base for base in allOtherBases]
mColors = [allColors.get(base, defaultColor) for base in allOtherBases]
# Add to listAnnt and listName
listAnnt.extend(mAnnts)
listName.extend(mNames)
listColor.extend(mColors)
# Add insertions
if 'i' in muttype:
# Make the list of annotations of the 4 possible insertions at the current position
iAnnts = [consensus+':0:1:0::+'+pos[1:]+base+'::' for base in allBases]
iName = ['+'+pos[1:]+base for base in allBases]
iColors = [allColors.get(base, defaultColor) for base in allBases]
# Add to listAnnt and listName
listAnnt.extend(iAnnts)
listName.extend(iName)
listColor.extend(iColors)
# Add deletions
if 'd' in muttype:
# Make annotation for the only possible deletion at the current position
dAnnt = consensus+':0:0:1:::'+pos+'x:'
dName = pos+'x'
dColor = allColors.get('x', defaultColor)
# Add to listAnnt and listName
listAnnt.append(dAnnt)
listName.append(dName)
listColor.append(dColor)
# Check if annotations exist in df, if not remove from listAnnt, listName, listColor, and xtickspos
if df.index.name == 'annotation':
listValidAnnt = [i for i, annt in enumerate(listAnnt) if annt in df.index]
else:
df2 = df.set_index('annotation')
listValidAnnt = [i for i, annt in enumerate(listAnnt) if annt in df2.index]
listAnnt = [item for i, item in enumerate(listAnnt) if i in listValidAnnt]
listName = [item for i, item in enumerate(listName) if i in listValidAnnt]
listColor = [item for i, item in enumerate(listColor) if i in listValidAnnt]
xtickspos = np.array([item for i, item in enumerate(xtickspos) if i in listValidAnnt])
# Call plotVariants to plot
if not colorbymut:
listColor = None
if fullname:
listName = None
if collapse:
xtickspos = None
return plotVariants(df, listAnnt, listName=listName, color=listColor, _xticks=xtickspos,
xticklabelrot=90, **kwargs)
def plotSingleMutants(df,
consensus,
listSeqPos,
muttype='m',
colorbymut=True,
fullname=False,
collapse=False,
**kwargs):
"""Plot the activities and counts of single mutants given a df of all clusters, name of consensus
sequence, and list of positions
This function defaults to plotting time/rate constants, but can be used to plot something else too"""
# Define constants
allBases = ['A', 'T', 'C', 'G']
allColors = {
'A': sns.xkcd_rgb["windows blue"],
'T': sns.xkcd_rgb["amber"],
'C': sns.xkcd_rgb["faded green"],
'G': sns.xkcd_rgb["reddish orange"],
'x': sns.xkcd_rgb["dusty purple"]
}
defaultColor = sns.xkcd_rgb["greyish"]
# Compute how many bars to be plotted at each position and xtickpos
numBarDict = {'m': 3, 'i': 3, 'd': 1}
numBar = np.sum([numBarDict.get(i, 0) for i in muttype])
gap = int(numBar) / 3
xtickspos = [
range((numBar + gap) * pos + 1, (numBar + gap) * pos + numBar + 1)
for pos in range(0, len(listSeqPos))
]
xtickspos = np.array([item for sublist in xtickspos for item in sublist])
# Make the list of annotations to plot
listAnnt = []
listName = []
listColor = []
for pos in listSeqPos:
# Add mismatches
if 'm' in muttype:
# Get base at the current position
currBase = pos[0]
# Get the 3 other bases at the current position
allOtherBases = seqlib.allOtherBases(currBase)
# Make the list of annotations of the 3 possible mismatches at the current position
mAnnts = [
consensus + ':1:0:0:' + pos + base + ':::'
for base in allOtherBases
]
mNames = [pos + base for base in allOtherBases]
mNames = [
name[:-1] if i % 3 == 1 else ' '
for i, name in enumerate(mNames)
]
mColors = [
allColors.get(base, defaultColor) for base in allOtherBases
]
# Add to listAnnt and listName
listAnnt.extend(mAnnts)
listName.extend(mNames)
listColor.extend(mColors)
# Add insertions
if 'i' in muttype:
# Make the list of annotations of the 4 possible insertions at the current position
iAnnts = [
consensus + ':0:1:0::+' + pos[1:] + base + '::'
for base in allBases
]
iName = ['+' + pos[1:] + base for base in allBases]
iColors = [allColors.get(base, defaultColor) for base in allBases]
# Add to listAnnt and listName
listAnnt.extend(iAnnts)
listName.extend(iName)
listColor.extend(iColors)
# Add deletions
if 'd' in muttype:
# Make annotation for the only possible deletion at the current position
dAnnt = consensus + ':0:0:1:::' + pos + 'x:'
dName = pos + 'x'
dColor = allColors.get('x', defaultColor)
# Add to listAnnt and listName
listAnnt.append(dAnnt)
listName.append(dName)
listColor.append(dColor)
# Check if annotations exist in df, if not remove from listAnnt, listName, listColor, and xtickspos
if df.index.name == 'annotation':
listValidAnnt = [
i for i, annt in enumerate(listAnnt) if annt in df.index
]
else:
df2 = df.set_index('annotation')
listValidAnnt = [
i for i, annt in enumerate(listAnnt) if annt in df2.index
]
listAnnt = [item for i, item in enumerate(listAnnt) if i in listValidAnnt]
listName = [item for i, item in enumerate(listName) if i in listValidAnnt]
listColor = [
item for i, item in enumerate(listColor) if i in listValidAnnt
]
xtickspos = np.array(
[item for i, item in enumerate(xtickspos) if i in listValidAnnt])
# Call plotVariants to plot
if not colorbymut:
listColor = None
if fullname:
listName = None
if collapse:
xtickspos = None
return plotVariants(df,
listAnnt,
listName=listName,
color=listColor,
_xticks=xtickspos,
xticklabelrot=90,
**kwargs)
def plotTertSMcoop(
dfUnqClusters,
listConsensus,
listSeqPos,
refConsensus,
field='params2.median',
listConsensusName=None,
vmin=None,
vmax=None,
cmap='RdBu',
c_bad='0.55',
robust=True,
figsize=None,
figunitheight=1,
figunitwidth=1,
maxfigwidth=32,
actLabel=r'$\mathrm{\mathsf{\Delta\Delta G^{\ddag}\ (kcal\ mol^{-1})}}$',
show=True,
**kwargs):
"""Plot tertiary contact-single point mutant cooperivity heatmap"""
# Define constants
R = 1.9872041e-3
T = 293.0
# Make dfUnqClusters indexed by annotation if not already
if dfUnqClusters.index.name == 'annotation':
dfUnqClusters2 = dfUnqClusters.copy()
else:
dfUnqClusters2 = dfUnqClusters.set_index('annotation')
# Make an empty dataframe for the TC-SPM matrix
mat_kobs = pd.DataFrame(columns=listConsensus)
mat_ddG = pd.DataFrame(columns=listConsensus)
# Make an empty series for the consensus variants
seriesConsensusAct = pd.Series(index=listConsensus)
# Make the list of annotations to plot
listConsensusAct = []
for con in listConsensus:
listAnnt = []
listName = []
for pos in listSeqPos:
# Get base at the current position
currBase = pos[0]
# Get the 3 other bases at the current position
allOtherBases = seqlib.allOtherBases(currBase)
# Make the list of annotations of the 3 possible mismatches at the current position
mAnnts = [
con + ':1:0:0:' + pos + base + ':::' for base in allOtherBases
]
mNames = [pos + base for base in allOtherBases]
# Add to listAnnt and listName
listAnnt.extend(mAnnts)
listName.extend(mNames)
# Get the activity of the consensus variants
seriesConsensusAct[con] = 1. / dfUnqClusters2.loc[
con + ':0:0:0::::'][field] * 60
# Add the list of activity of the annotations from the current consensus as a column
mat_kobs[con] = (1. / dfUnqClusters2.loc[listAnnt][field] *
60).tolist()
# Compute delta delta G from k_obs
for con in listConsensus:
mat_ddG[con] = -R * T * np.log(
(mat_kobs[con] * seriesConsensusAct[refConsensus]) /
(mat_kobs[refConsensus] * seriesConsensusAct[con]))
#
if listConsensusName is not None:
mat_kobs.columns = listConsensusName
mat_ddG.columns = listConsensusName
#
mat_kobs.columns.name = 'Tertiary contact knockouts'
mat_ddG.columns.name = 'Tertiary contact knockouts'
# Set single mutant names
mat_kobs['Single point mutants'] = listName
mat_kobs = mat_kobs.set_index('Single point mutants').transpose()
mat_ddG['Single point mutants'] = listName
mat_ddG = mat_ddG.set_index('Single point mutants').transpose()
# Define colormap
if isinstance(cmap, basestring):
cmap = plt.get_cmap(cmap)
cmap.set_bad(c_bad, 0.8)
# Make mask for nan data
mask = ~np.isfinite(mat_ddG)
# Plot heatmap
if not figsize:
figwidth = min(figunitwidth * len(listAnnt), maxfigwidth)
figheight = figunitheight * len(listConsensus) + 0.5
figsize = (figwidth, figheight)
fig, ax = plt.subplots(1, 1, figsize=figsize)
cbar_ax = fig.add_axes([.78, .83, .2, .05])
sns.heatmap(mat_ddG,
ax=ax,
square=True,
mask=mask,
robust=robust,
vmin=vmin,
vmax=vmax,
center=0,
cmap=cmap,
cbar_ax=cbar_ax,
cbar_kws={'orientation': 'horizontal'})
setproperties(ax=cbar_ax, fontsize=21, xlabel=actLabel)
setproperties(ax=ax,
tickfontsize=22,
yticklabelrot=0,
labelfontsize=28,
tight=True,
pad=1.0)
if show:
plt.show(block=False)
return