def show_rna_structs(xvals, yvals, structs, energies, pfracs,
rname, rtype,ns,rfid,figsize,colors,
seq, n, selection_type,vert_idxs):
verts = rutils.struct_verts([structs['structs'][i] for i in vert_idxs]
,seq,rfid)
f = myplots.fignum(3,figsize)
ax = f.add_subplot(111)
myplots.padded_limits(ax, xvals, yvals, .2)
for vi, v in enumerate(verts):
i = vert_idxs[vi]
dims = [30]
shadow_width = 10
pkw0 = {'linewidth':shadow_width,
'color':'white',
'alpha':1,
'zorder':1.1}
rplots.show_rna([xvals[i],yvals[i]], v,
dims = dims,
pkw = pkw0)
pkw0 = {'linewidth':shadow_width,
'color':'white',
'alpha':.8,
'zorder':vi+2}
rplots.show_rna([xvals[i],yvals[i]], v,
dims = dims,
pkw = pkw0)
pkw1 = {'linewidth':2,
'color':colors[i],
'zorder':vi+2}
rplots.show_rna([xvals[i],yvals[i]], v,
dims = dims, pkw = pkw1)
ax.set_ylabel('mutation score')
ax.set_xlabel('free energy (-kCal)')
ax.annotate('''Suboptimal foldings, positioned by energy and
a mutation based evolutionary score.
Color indicates a second score from paired BL.''' , [0,1],xycoords ='axes fraction',
xytext = [10,-10], textcoords='offset pixels',
va = 'top')
f.savefig(figfile.format('{3}_frac_silent_{0}_{1}{2}'.\
format(rname,selection_type,n,rtype)))
return vert_idxs
def show_rna_structs(xvals, yvals, structs, energies, pfracs, rname, rtype, ns,
rfid, figsize, colors, seq, n, selection_type, vert_idxs):
verts = rutils.struct_verts([structs['structs'][i] for i in vert_idxs],
seq, rfid)
f = myplots.fignum(3, figsize)
ax = f.add_subplot(111)
myplots.padded_limits(ax, xvals, yvals, .2)
for vi, v in enumerate(verts):
i = vert_idxs[vi]
dims = [30]
shadow_width = 10
pkw0 = {
'linewidth': shadow_width,
'color': 'white',
'alpha': 1,
'zorder': 1.1
}
rplots.show_rna([xvals[i], yvals[i]], v, dims=dims, pkw=pkw0)
pkw0 = {
'linewidth': shadow_width,
'color': 'white',
'alpha': .8,
'zorder': vi + 2
}
rplots.show_rna([xvals[i], yvals[i]], v, dims=dims, pkw=pkw0)
pkw1 = {'linewidth': 2, 'color': colors[i], 'zorder': vi + 2}
rplots.show_rna([xvals[i], yvals[i]], v, dims=dims, pkw=pkw1)
ax.set_ylabel('mutation score')
ax.set_xlabel('free energy (-kCal)')
ax.annotate('''Suboptimal foldings, positioned by energy and
a mutation based evolutionary score.
Color indicates a second score from paired BL.''', [0, 1],
xycoords='axes fraction',
xytext=[10, -10],
textcoords='offset pixels',
va='top')
f.savefig(figfile.format('{3}_frac_silent_{0}_{1}{2}'.\
format(rname,selection_type,n,rtype)))
return vert_idxs
def show_conservation(fidx = 0, reset = False):
fnum = flist[fidx]
rfid = 'RF{0:05}'.format(fnum)
print rfid
if fnum ==50: ftype = 'riboswitch'
else: ftype = 'all'
out = mem.getOrSet(setFamData,
**mem.rc({}, reset =reset,
on_fail = 'compute',
hardcopy = False,
register = 'fdat'+rfid,
ftype = ftype,
rfid = rfid))
mvals, tvals, structs = mem.getOrSet(setTree,
**mem.rc({},reset = reset,
on_fail = 'compute',
hardcopy = True,
register = 'st'+rfid,
rfid = rfid,
ftype = ftype))
idxs, tidx = sutils.show_paired_v_energy(rfid,rfid,mvals,tvals,structs,ftype)
all_pairs = structs['structs']
all_energies = structs['energies']
pints,eints, mints, tints = [structs['structs'][i] for i in idxs],\
[ structs['energies'][i] for i in idxs],\
[ mvals[tidx][i] for i in idxs],\
[ tvals[tidx][i] for i in idxs]
seq = structs['seq']
if do_make_subopts:
subopts = rutils.suboptimals(seq, n = 400)
verts = rutils.struct_verts(subopts, seq, rfid)
f = myplots.fignum(4,figsize)
rplots.grid_rnas(verts, dims = [40])
f.savefig(figfile.format('{0}_grid_rnas'.\
format(rfid)))
aff = rutils.struct_affinity_matrix(all_pairs, len(seq))
pca = rutils.project_structs(all_pairs,
ptype ='pca',
affinities = aff,
n_comp = 3)
for metric in ['n_comp']:# ['frac_silent','frac_paired','n_comp']:
scolors = []
for i in range(len(tvals[tidx])):
m_silent, pidxs, frac_good = sutils.metric(
mvals[tidx][i],tvals[tidx][i],
mtype = metric)
scolors.append(mean(m_silent))
scolors = myplots.rescale(scolors, [0.,1.])[:,newaxis] * array([1.,0.,0.])
f = myplots.fignum(4,figsize)
ax = f.add_subplot(111)
xvals, yvals = pca[:,:2].T
myplots.padded_limits(ax, xvals, yvals)
ax.scatter(xvals,yvals,300,linewidth = 1,
edgecolor = 'black', color = scolors)
ax.scatter(pca[idxs,0],pca[idxs,1], 2100 ,alpha = 1,
color = 'black')
ax.scatter(pca[idxs,0],pca[idxs,1], 2000 ,alpha = 1,
color = 'white')
ax.scatter(pca[idxs,0],pca[idxs,1], 400 ,alpha = 1,
color = scolors[idxs],
)
ax.annotate('''Conservation metric: {0}
Projected onto C=2 Principal Components'''.format(metric),
[0,1],xycoords = 'axes fraction', va = 'top',
xytext = [10,-10],textcoords='offset points')
f.savefig(figfile.format('{0}_pca_{1}'.\
format(rfid, metric)))
def show_paired_v_energy(rname,rfid, all_muts, all_times, structs,rtype):
if all_times == {}:
return
resolved_frac = [ mean(list(it.chain(*[s_times['frac_resolved']
for s_times in t_times.values()])))
for t_times in all_times.values()]
total_lens = [mean(list(it.chain(*[s_times['total_time']
for s_times in t_times.values()])))
for t_times in all_times.values()]
total_lens_res = [mean(list(it.chain(*[s_times['total_time_res']
for s_times in t_times.values()])))
for t_times in all_times.values()]
focus_tree = all_times.keys()[argmax(total_lens_res)]
muts = all_muts[focus_tree]
times = all_times[focus_tree]
ns = len(muts.keys())
s2 = dict(structs)
s2['energies'] = s2['energies'][:ns]
s2['structs'] = s2['structs'][:ns]
structs = s2
energies = structs['energies']
f = myplots.fignum(3,figsize)
xvals, yvals , pfracs, ugfracs = [], [], [], []
for i, vals in enumerate(zip(muts.values(),times.values())):
mvals, tvals = vals
xvals.append( energies[i])
frac_ug = metric(mvals, tvals, 'frac_ug')[0]
pfrac,pinds,frac_good = metric(mvals, tvals, 'frac_silent')
sfrac = metric(mvals, tvals, 'frac_silent')[0]
ugfracs.append(frac_ug)
pfracs.append( mean(pfrac))
yvals.append( mean(sfrac)*frac_good)
colors = array(pfracs)
colors = (colors - min(colors)) /(max(colors) - min(colors))
colors = colors[:,newaxis] * [0,1,0]
ax = f.add_subplot(111)
ax.scatter(xvals,yvals,array(ugfracs) * 200,
color = colors)
ax.set_ylabel('mutation score')
ax.set_xlabel('free energy (-kCal)')
ax.annotate('''Evaluated structures positioned by energy
and a mutation based evolutionary score.
Color indicates fractional frequency of double mutants.
Radius indicates percentage of ungapped base pairs.''' , [0,1],xycoords = 'axes fraction',
xytext = [10,-10], textcoords='offset pixels',
va = 'top')
myplots.padded_limits(ax, xvals, yvals, .2)
f.savefig(figfile.format('{1}_frac_double_{0}'.format(rname,rtype)))
f.clear()
colors = array(pfracs)
colors = (colors - min(colors)) /(max(colors) - min(colors))
colors = colors[:,newaxis] * [1,0,0]
f = myplots.fignum(3,figsize)
xvals, yvals , pfracs, ugfracs = [], [], [], []
for i, vals in enumerate(zip(muts.values(),times.values())):
mvals, tvals = vals
xvals.append( energies[i])
frac_ug = metric(mvals, tvals, 'frac_ug')[0]
pfrac,pinds,frac_good = metric(mvals, tvals, 'frac_paired')
sfrac = metric(mvals, tvals, 'frac_silent')[0]
ugfracs.append(frac_ug)
pfracs.append( mean(pfrac)*frac_good)
yvals.append( mean(sfrac)*frac_good)
colors = array(pfracs)
colors = (colors - min(colors)) /(max(colors) - min(colors))
colors = colors[:,newaxis] * [1,0,0]
ax = f.add_subplot(111)
ax.scatter(xvals,yvals,array(ugfracs) * 200,
color = colors)
ax.set_ylabel('mutation score')
ax.set_xlabel('free energy (-kCal)')
ax.annotate('''Evaluated structures positioned by energy
and a mutation based evolutionary score.
Color indicates a second score from paired BL.
Radius indicates percentage of ungapped base pairs.''' , [0,1],
xycoords = 'axes fraction',
xytext = [10,-10], textcoords='offset pixels',
va = 'top')
myplots.padded_limits(ax, xvals, yvals, .2)
f.savefig(figfile.format('{1}_frac_silent_{0}'.format(rname,rtype)))
seq = structs['seq']
n, selection_type = [4,'both']
idxs = get_interesting_inds(xvals, yvals, structs, energies, pfracs,
rname, rtype, ns,rfid,figsize, colors,
seq,n,selection_type)
if draw_single:
show_rna_structs(xvals, yvals, structs, energies, pfracs,
rname, rtype, ns,rfid,figsize, colors,
seq,n,selection_type, idxs)
if draw_many:
for n, selection_type in \
[[5,'ptime'],[5,'energy'],[ns,'energy']]:
m_idxs = get_interesting_inds(xvals, yvals, structs, energies, pfracs,
rname, rtype, ns,rfid,figsize, colors,
seq,n,selection_type)
show_rna_structs(xvals, yvals, structs, energies, pfracs,
rname, rtype, ns,rfid,figsize, colors,
seq,n,selection_type, m_idxs)
return idxs,focus_tree
def show_paired_v_energy(rname, rfid, all_muts, all_times, structs, rtype):
if all_times == {}:
return
resolved_frac = [
mean(
list(
it.chain(
*
[s_times['frac_resolved']
for s_times in t_times.values()])))
for t_times in all_times.values()
]
total_lens = [
mean(
list(
it.chain(
*[s_times['total_time'] for s_times in t_times.values()])))
for t_times in all_times.values()
]
total_lens_res = [
mean(
list(
it.chain(*[
s_times['total_time_res'] for s_times in t_times.values()
]))) for t_times in all_times.values()
]
focus_tree = all_times.keys()[argmax(total_lens_res)]
muts = all_muts[focus_tree]
times = all_times[focus_tree]
ns = len(muts.keys())
s2 = dict(structs)
s2['energies'] = s2['energies'][:ns]
s2['structs'] = s2['structs'][:ns]
structs = s2
energies = structs['energies']
f = myplots.fignum(3, figsize)
xvals, yvals, pfracs, ugfracs = [], [], [], []
for i, vals in enumerate(zip(muts.values(), times.values())):
mvals, tvals = vals
xvals.append(energies[i])
frac_ug = metric(mvals, tvals, 'frac_ug')[0]
pfrac, pinds, frac_good = metric(mvals, tvals, 'frac_silent')
sfrac = metric(mvals, tvals, 'frac_silent')[0]
ugfracs.append(frac_ug)
pfracs.append(mean(pfrac))
yvals.append(mean(sfrac) * frac_good)
colors = array(pfracs)
colors = (colors - min(colors)) / (max(colors) - min(colors))
colors = colors[:, newaxis] * [0, 1, 0]
ax = f.add_subplot(111)
ax.scatter(xvals, yvals, array(ugfracs) * 200, color=colors)
ax.set_ylabel('mutation score')
ax.set_xlabel('free energy (-kCal)')
ax.annotate('''Evaluated structures positioned by energy
and a mutation based evolutionary score.
Color indicates fractional frequency of double mutants.
Radius indicates percentage of ungapped base pairs.''', [0, 1],
xycoords='axes fraction',
xytext=[10, -10],
textcoords='offset pixels',
va='top')
myplots.padded_limits(ax, xvals, yvals, .2)
f.savefig(figfile.format('{1}_frac_double_{0}'.format(rname, rtype)))
f.clear()
colors = array(pfracs)
colors = (colors - min(colors)) / (max(colors) - min(colors))
colors = colors[:, newaxis] * [1, 0, 0]
f = myplots.fignum(3, figsize)
xvals, yvals, pfracs, ugfracs = [], [], [], []
for i, vals in enumerate(zip(muts.values(), times.values())):
mvals, tvals = vals
xvals.append(energies[i])
frac_ug = metric(mvals, tvals, 'frac_ug')[0]
pfrac, pinds, frac_good = metric(mvals, tvals, 'frac_paired')
sfrac = metric(mvals, tvals, 'frac_silent')[0]
ugfracs.append(frac_ug)
pfracs.append(mean(pfrac) * frac_good)
yvals.append(mean(sfrac) * frac_good)
colors = array(pfracs)
colors = (colors - min(colors)) / (max(colors) - min(colors))
colors = colors[:, newaxis] * [1, 0, 0]
ax = f.add_subplot(111)
ax.scatter(xvals, yvals, array(ugfracs) * 200, color=colors)
ax.set_ylabel('mutation score')
ax.set_xlabel('free energy (-kCal)')
ax.annotate('''Evaluated structures positioned by energy
and a mutation based evolutionary score.
Color indicates a second score from paired BL.
Radius indicates percentage of ungapped base pairs.''', [0, 1],
xycoords='axes fraction',
xytext=[10, -10],
textcoords='offset pixels',
va='top')
myplots.padded_limits(ax, xvals, yvals, .2)
f.savefig(figfile.format('{1}_frac_silent_{0}'.format(rname, rtype)))
seq = structs['seq']
n, selection_type = [4, 'both']
idxs = get_interesting_inds(xvals, yvals, structs, energies, pfracs, rname,
rtype, ns, rfid, figsize, colors, seq, n,
selection_type)
if draw_single:
show_rna_structs(xvals, yvals, structs, energies, pfracs, rname, rtype,
ns, rfid, figsize, colors, seq, n, selection_type,
idxs)
if draw_many:
for n, selection_type in \
[[5,'ptime'],[5,'energy'],[ns,'energy']]:
m_idxs = get_interesting_inds(xvals, yvals, structs, energies,
pfracs, rname, rtype, ns, rfid,
figsize, colors, seq, n,
selection_type)
show_rna_structs(xvals, yvals, structs, energies, pfracs, rname,
rtype, ns, rfid, figsize, colors, seq, n,
selection_type, m_idxs)
return idxs, focus_tree
