def top6split(run, x, k, n):
values = [*khams[run][k][n][x].keys()]
vals = []
if n == 1:
top6s = top6all[run][k].copy()
else:
top6s = top6all[run][k].copy()
for j in top6s:
if j in removelist[run][k][n]:
top6s.remove(j)
rj = kmer2hash(revComp(hash2kmer(j, k)))
if rj in top6s:
top6s.remove(rj)
for i in top6s[::2]:
if i in values:
vals.append(i)
rk = kmer2hash(revComp(hash2kmer(i, k)))
vals.append(rk)
try:
if rk not in khams[run][k][n][x]:
khams[run][k][n][x][rk] = 0
except:
continue
return vals
def scatter(run, k):
top6(run, k)
split = top6splitter(run, k)
labels = []
handels = []
labelcolours = []
#greys = [0.6, 0.75, 0.9]
c = 0
cc = 0
for x in khams[run][k]:
xaxis = xaxismaker(run, x, k)
yaxis = yaxismaker(run, x, k)
top6x = top6plotx(run, x, k)
top6y = top6ploty(run, x, k)
plt.scatter(xaxis, yaxis, color = '0.75', alpha=0.7, s=1)
count = len(top6x)
for i in range(0, count):
plt.scatter(top6x[i], top6y[i], label=split[run][x][i], color=colours[c], s=3)
labels.append((str(hash2kmer(split[run][x][i],k)+' '+str(khams[run][k][x][(split[run][x][i])]))))
c += 1
for p, j in enumerate(split[run][x]):
xp = top6x[p]
yp = top6y[p]
if (p % 2)+x == x:
plt.annotate(str(hash2kmer(j,k) + ' ' + str(khams[run][k][x][j])), (float(xp), float(yp)+0.6), color=colours[cc], fontsize=5)
if (p % 2)+x != x:
plt.annotate(str(hash2kmer(j,k) + ' ' + str(khams[run][k][x][j])), (float(xp), float(yp)-0.6), color=colours[cc], fontsize=5)
cc += 1
for z in colours[::2]:
labelcolours.append(z)
leg = plt.legend(labels[::2], fontsize=7)
for i in range(0,6):
leg.legendHandles[i].set_color(labelcolours[i])
leg.legendHandles[i]._sizes = [8]
plt.xlabel("Hamming distance")
plt.ylabel("Kmer count")
plt.title("Run number:"+' '+str(run)+'\n'+'K: '+str(k)+'\n'+'Total kmers: '+str(len(runlists[run][k])))
#plt.show()
plt.savefig('figures/hamdist_'+str(identifier)+"_"+str(run)+"_"+str(k), dpi=600)
plt.close()
def hammer(k):
for _ in range(numofruns + 1):
khams.append({})
for j in range(1,len(kmercount)):
khams[j][k] = {k: {} for k in range(k+1)}
hconsensus = max(kmercount[j][k], key=lambda key: kmercount[j][k][key])
consensus = hash2kmer(hconsensus, k)
for x in list(kmercount[j][k].keys()):
values = hash2kmer(x,k)
rvalues = revComp(values)
ham = hamming_distance(consensus, values)
rham = hamming_distance(consensus, rvalues)
if ham <= rham:
khams[j][k][ham].update({x:kmercount[j][k][x]})
if ham > rham:
khams[j][k][rham].update({x:kmercount[j][k][x]})
return khams
def makeyaxis2c(i, k):
yaxis2c = ([])
rkmer = kmer2hash(revComp(hash2kmer(i, k)))
top6s = top6(k)
for l in range(1, (len(kmercount))):
total = sum(kmercount[l][k].values())
if i in top6s:
num = kmercount[l][k][i]
rnum = kmercount[l][k][rkmer]
yaxis2c.append((num + rnum) / total)
return yaxis2c
def makeyaxis2b(i, k):
yaxis2b = ([])
rkmer = kmer2hash(revComp(hash2kmer(i, k)))
colours = colours1(k)
for l in range(1, (len(kmercount))):
total = sum(kmercount[l][k].values())
if i in colours:
num = kmercount[l][k][i]
rnum = kmercount[l][k][rkmer]
yaxis2b.append((num + rnum) / total)
return yaxis2b
def makeyaxis1c(i, k):
yaxis1c = ([])
rkmer = kmer2hash(revComp(hash2kmer(i, k)))
top6s = top6(k)
for l in range(1, (len(kmercount))):
total = sum(kmercount[l][k].values())
if i in top6s:
num = kmercount[l][k][i]
rnum = kmercount[l][k][rkmer]
f = (num + rnum) / total
freq = (f / (1 - f))
yaxis1c.append(math.log10(freq))
return yaxis1c
def hammer():
for _ in range(numofruns + 1):
khams.append({})
for run in range(1, numofruns + 1):
for k in range(mink, maxk + 1):
khams[run][k] = {}
for n in range(1, nmotifs + 1):
khams[run][k][n] = {k: {} for k in range(k + 1)}
#print("run "+str(run)+" k "+str(k))
#print(khams)
if n == 1:
hconsensus = max(kmercount[run][k],
key=lambda key: kmercount[run][k][key])
else:
temptop6 = top6all[run][k].copy()
for j in temptop6:
if j in removelist[run][k][n]:
temptop6.remove(j)
if kmer2hash(revComp(hash2kmer(j, k))) in temptop6:
temptop6.remove(
kmer2hash(revComp(hash2kmer(j, k))))
hconsensus = max(temptop6,
key=lambda key: kmercount[run][k][key])
consensus = hash2kmer(hconsensus, k)
#print("Consensus: "+str(consensus))
for x in list(kmercount[run][k].keys()):
if x not in removelist[run][k][n]:
values = hash2kmer(x, k)
rvalues = revComp(values)
ham = hamming_distance(consensus, values)
rham = hamming_distance(consensus, rvalues)
if ham <= rham:
khams[run][k][n][ham].update(
{x: kmercount[run][k][x]})
if ham > rham:
khams[run][k][n][rham].update(
{x: kmercount[run][k][x]})
return khams
def top6(run, k):
top6s = []
keys = list(kmercount[run][k].keys())
topvalpos = 0
while len(top6s) <= 11:
next = keys[topvalpos]
nkmer = hash2kmer(next, k)
nrkmer = revComp(nkmer)
nhrkmer = kmer2hash(nrkmer)
if next not in top6s:
top6s.append(next)
if nhrkmer not in top6s:
top6s.append(nhrkmer)
topvalpos += 1
return top6s
def makeyaxis2b(i, k):
yaxis2b = ([])
rkmer = kmer2hash(revComp(hash2kmer(i, k)))
for l in range(1, numofruns + 1):
total = totaldict[l][k]
if i in colours[k]:
try:
num = kmercount[l][k][i]
except:
num = 0
try:
rnum = kmercount[l][k][rkmer]
except:
rnum = 0
yaxis2b.append((num + rnum) / total)
return yaxis2b
def makeyaxis2c(i, k):
yaxis2c = ([])
rkmer = kmer2hash(revComp(hash2kmer(i, k)))
top6s = top6all[numofruns][k]
for l in range(1, numofruns + 1):
total = totaldict[l][k]
if i in top6s:
try:
num = kmercount[l][k][i]
except:
num = 0
try:
rnum = kmercount[l][k][rkmer]
except:
rnum = 0
yaxis2c.append((num + rnum) / total)
return yaxis2c
def makeyaxis1b(i, k):
yaxis1b = ([])
rkmer = kmer2hash(revComp(hash2kmer(i, k)))
for l in range(1, numofruns + 1):
total = totaldict[l][k]
if i in colours[k]:
try:
num = kmercount[l][k][i]
except:
num = 0
try:
rnum = kmercount[l][k][rkmer]
except:
rnum = 0
f = (num + rnum) / total
freq = (f / (1 - f))
yaxis1b.append(math.log10(freq))
return yaxis1b
def makeyaxis1c(i, k):
yaxis1c = ([])
rkmer = kmer2hash(revComp(hash2kmer(i, k)))
top6s = top6all[numofruns][k]
for l in range(1, numofruns + 1):
total = totaldict[l][k]
if i in top6s:
try:
num = kmercount[l][k][i]
except:
num = 0
try:
rnum = kmercount[l][k][rkmer]
except:
rnum = 0
f = (num + rnum) / total
freq = (f / (1 - f))
yaxis1c.append(math.log10(freq))
return yaxis1c
def xaxismaker(run, p, k, n):
x = []
keys = [*khams[run][k][n][p].keys()]
if n == 1:
top6s = top6all[run][k].copy()
else:
top6s = top6all[run][k].copy()
for j in top6s:
if j in removelist[run][k][n]:
top6s.remove(j)
rj = kmer2hash(revComp(hash2kmer(j, k)))
if rj in top6s:
top6s.remove(rj)
for i in top6s:
if i in keys:
keys.remove(i)
size = len(keys)
for _ in range(size):
nval = float(p) + random.uniform(-0.3, 0.3)
x.append(nval)
return x
def yaxismaker(run, x, k, n):
y = []
values = [*khams[run][k][n][x].keys()]
if n == 1:
top6s = top6all[run][k].copy()
else:
top6s = top6all[run][k].copy()
for j in top6s:
if j in removelist[run][k][n]:
top6s.remove(j)
rj = kmer2hash(revComp(hash2kmer(j, k)))
if rj in top6s:
top6s.remove(rj)
for i in top6s:
if i in values:
values.remove(i)
for i in values:
val = khams[run][k][n][x][i]
nval = float(val) + random.uniform(-0.1, 0.1)
y.append(nval)
return y
def top6plotx(run, p, k, n):
x = []
keys = [*khams[run][k][n][p].keys()]
vals = []
if n == 1:
top6s = top6all[run][k].copy()
else:
top6s = top6all[run][k].copy()
for j in top6s:
if j in removelist[run][k][n]:
top6s.remove(j)
rj = kmer2hash(revComp(hash2kmer(j, k)))
if rj in top6s:
top6s.remove(rj)
#print("TOP6Sx")
#print(top6s)
for i in top6s:
if i in keys:
vals.append(i)
for _ in range(len(vals)):
nval = float(p) + random.uniform(-0.3, 0.3)
x.append(nval)
return x
def grapher(k):
xaxis = makexaxis()
last = (len(xaxis) - 1)
top6s = top6all[numofruns][k]
fig = plt.figure(figsize=(10, 10))
grid = plt.GridSpec(2, 3, wspace=0.4, hspace=0.3)
top = fig.add_subplot(grid[:-1, :])
top.set_xlabel("SELEX round")
top.set_ylabel("log(f/(1-f))")
top.set_title("Kmer frequency" + ', K: ' + str(k))
top.set_xlim([((startround - 1) - 0.5), (numofruns + startround)])
top.set_xticks(
np.linspace((startround - 1) - 0.5, (numofruns + startround),
num=((2 * (numofruns + 1)) + 2),
endpoint=True))
bottom = fig.add_subplot(grid[-1, :-1])
bottom.set_xlabel("SELEX round")
bottom.set_ylabel("f = (kmer/total)")
bottom.set_title("Kmer frequency")
bottom.set_xlim([((startround - 1) - 0.5), (numofruns + (startround - 1))])
bottom.set_xticks(range((startround - 1), (numofruns + 1)))
bar = fig.add_subplot(grid[-1, -1:])
bar.set_xlabel("SELEX round")
bar.set_ylabel("Total kmers")
bar.set_title("Kmer total distribution")
bar.set_xticks(makexaxis())
colourslist = ['C0', 'C1', 'C2', 'C3', 'C4', 'C5', 'C6', 'C7']
if revcompwanted == False:
for l in range(1, (numofruns + 1)):
for i in kmercount[l][k]:
num = kmercount[l][k][i]
rkmer = kmer2hash(revComp(hash2kmer(i, k)))
try:
rnum = kmercount[l][k][rkmer]
if num > rnum:
if i in top6s:
yaxis1c = makeyaxis1c(i, k)
top.plot(xaxis,
yaxis1c,
color=colourslist[(top6s.index(i) // 2)],
linewidth=2,
marker="s",
markevery=None,
zorder=int(kmercount[l][k][i]))
if i in colours[k]:
yaxis1b = makeyaxis1b(i, k)
top.plot(xaxis,
yaxis1b,
linewidth=1,
marker=".",
markevery=None,
zorder=int(kmercount[l][k][i]))
if i not in top6s or colours[k]:
yaxis1a = makeyaxis1a(i, k)
top.plot(xaxis,
yaxis1a,
color='0.75',
linestyle='--',
linewidth=0.5,
marker="x",
markevery=None,
alpha=0.5,
zorder=0)
except:
continue
for l in range(1, (numofruns + 1)):
for i in kmercount[l][k]:
num = kmercount[l][k][i]
rkmer = kmer2hash(revComp(hash2kmer(i, k)))
try:
rnum = kmercount[l][k][rkmer]
if num > rnum:
if i in top6s:
yaxis2c = makeyaxis2c(i, k)
bottom.plot(xaxis,
yaxis2c,
color=colourslist[(top6s.index(i) //
2)],
linewidth=2,
marker="s",
markevery=None,
zorder=int(kmercount[l][k][i]))
if i in colours[k]:
yaxis2b = makeyaxis2b(i, k)
bottom.plot(xaxis,
yaxis2b,
linewidth=1,
marker=".",
markevery=None,
zorder=int(kmercount[l][k][i]))
if i not in top6s or colours[k]:
yaxis2a = makeyaxis2a(i, k)
bottom.plot(xaxis,
yaxis2a,
color='0.75',
linestyle='--',
linewidth=0.5,
marker="x",
markevery=None,
alpha=0.8,
zorder=0)
except:
continue
if revcompwanted == True:
for l in range(1, (numofruns + 1)):
alreadytop = []
for i in kmercount[l][k]:
num = kmercount[l][k][i]
rkmer = kmer2hash(revComp(hash2kmer(i, k)))
try:
rnum = kmercount[l][k][rkmer]
if num >= rnum:
alreadytop.append(rkmer)
if i not in alreadytop:
if i in top6s:
yaxis1c = makeyaxis1c(i, k)
top.plot(xaxis,
yaxis1c,
color=colourslist[(top6s.index(i) //
2)],
linewidth=2,
marker="s",
markevery=None,
zorder=int(kmercount[l][k][i]))
if i in colours[k]:
yaxis1b = makeyaxis1b(i, k)
top.plot(xaxis,
yaxis1b,
linewidth=1,
marker=".",
markevery=None,
zorder=int(kmercount[l][k][i]))
if i not in top6s or colours[k]:
yaxis1a = makeyaxis1a(i, k)
top.plot(xaxis,
yaxis1a,
color='0.75',
linestyle='--',
linewidth=0.5,
marker="x",
markevery=None,
alpha=0.5,
zorder=0)
except:
continue
for l in range(1, numofruns + 1):
alreadybottom = []
for i in kmercount[l][k]:
num = kmercount[l][k][i]
rkmer = kmer2hash(revComp(hash2kmer(i, k)))
try:
rnum = kmercount[l][k][rkmer]
if num >= rnum:
alreadybottom.append(rkmer)
if i not in alreadybottom:
if i in top6s:
yaxis2c = makeyaxis2c(i, k)
bottom.plot(
xaxis,
yaxis2c,
color=colourslist[(top6s.index(i) // 2)],
linewidth=2,
marker="s",
markevery=None,
zorder=int(kmercount[l][k][i]))
if i in colours[k]:
yaxis2b = makeyaxis2b(i, k)
bottom.plot(xaxis,
yaxis2b,
linewidth=1,
marker=".",
markevery=None,
zorder=int(kmercount[l][k][i]))
if i not in top6s or colours[k]:
yaxis2a = makeyaxis2a(i, k)
bottom.plot(xaxis,
yaxis2a,
color='0.75',
linestyle='--',
linewidth=0.5,
marker="x",
markevery=None,
alpha=0.8,
zorder=0)
except:
continue
ymint, ymaxt = top.get_ylim()
ypost = np.linspace(ymint, ymaxt, num=20, endpoint=True)
yminb, ymaxb = bottom.get_ylim()
yposb = np.linspace(yminb, ymaxb, num=20, endpoint=True)
top6labels = []
for n, i in enumerate(top6s[::2]):
p = (n // 2)
yaxis1c = makeyaxis1c(i, k)
yaxis2c = makeyaxis2c(i, k)
top.annotate((str(n + 1)), (xaxis[last], yaxis1c[last]),
(xaxis[last] + 0.2, ypost[-(n + 2)]),
size=10,
fontname='monospace',
weight='bold',
arrowprops=dict(color=colourslist[n],
shrink=0.05,
width=0.05,
headwidth=0.4),
color=colourslist[n])
top6labels.append(
str(n + 1) + ". " + str(hash2kmer(i, k)) + " / " +
revComp(str(hash2kmer(i, k))))
bottom.annotate((str(n + 1) + ". " + str(hash2kmer(i, k))),
(xaxis[last], yaxis2c[last]),
(xaxis[last] + 0.3, yposb[-(n + 2)]),
size=10,
fontname='monospace',
weight='bold',
color=colourslist[n],
arrowprops=dict(color=colourslist[n],
shrink=0.05,
width=0.05,
headwidth=0.4))
for p, i in enumerate(top6s[::2]):
dp = (p + 3)
top.text(s=(str(top6labels[p])),
x=(numofruns - 0.5 + (startround - 1)),
y=(ypost[-dp]),
size=14,
fontname='monospace',
color=colourslist[p])
bar.bar(xaxis, makeyaxis3(k))
plt.savefig("figures/" + str(identifier) + "/kmer_frequency/kmerfreq_" +
str(identifier) + "_" + str(k),
dpi=600)
plt.close()
def scatter(run, k, n):
#top6(run, k)
split = top6splitter(run, k, n)
#print("split")
#print(split)
labels = []
handels = []
c = 0
cc = 0
texts = []
for x in khams[run][k][n]:
xaxis = xaxismaker(run, x, k, n)
yaxis = yaxismaker(run, x, k, n)
top6x = top6plotx(run, x, k, n)
#print("top6x")
#print(top6x)
top6y = top6ploty(run, x, k, n)
#print("top6y")
#print(top6y)
plt.scatter(xaxis, yaxis, color='0.75', alpha=0.7, s=1)
count = len(top6x)
for i in range(0, count):
plt.scatter(top6x[i],
top6y[i],
label=split[run][x][i],
color=colours[c],
s=3)
labels.append((str(
hash2kmer(split[run][x][i], k) + ' ' +
str(khams[run][k][n][x][(split[run][x][i])]))))
c += 1
for p, j in enumerate(split[run][x]):
xp = top6x[p]
yp = top6y[p]
texts.append(
plt.text(
float(xp),
float(yp - 0.3),
str(hash2kmer(j, k) + ' ' + str(khams[run][k][n][x][j])),
color=colours[cc],
fontsize=5))
cc += 1
adjust_text(texts, lw=0.5)
leg = plt.legend(labels[::2], fontsize=7)
if n == 1:
top6s = top6all[run][k].copy()
else:
top6s = top6all[run][k].copy()
for j in top6s:
if j in removelist[run][k][n]:
top6s.remove(j)
rj = kmer2hash(revComp(hash2kmer(j, k)))
if rj in top6s:
top6s.remove(rj)
for i in range(0, (len(top6s) // 2)):
leg.legendHandles[i].set_color(labelcolours[i])
leg.legendHandles[i]._sizes = [8]
plt.xticks(np.arange(0, len(khams[run][k][n]), step=1))
plt.xlabel("Hamming distance")
plt.ylabel("Kmer count")
if n == 1:
plt.title("Run number:" + ' ' + str(run +
(startround - 1)) + '\n' + 'K: ' +
str(k) + '\n' + 'Total kmers: ' + str(totaldict[run][k]))
else:
plt.title("Run number:" + ' ' + str(run + (startround - 1)) + '\n' +
'K: ' + str(k) + '\n' + 'Total kmers: ' +
str(totaldict[run][k] - removedkmers[run][k][n]))
plt.savefig("figures/" + str(identifier) + "/hamming_distance/hamdist_" +
str(identifier) + "_" + str(run + (startround - 1)) + "_" +
str(k) + "_" + str(n),
dpi=600)
plt.close()
