def pwmmaker():
for runnum in range(1, numofruns+1):
for i in hamdict[runnum]:
for n in range(1, nmotifs+1):
for x in hamdict[runnum][i][n]:
kmer = hash2kmer((x),i)
for j in range(1,i+1):
if kmer[j-1] == "A":
pwm[runnum][i][n][j]["A"] += hamdict[runnum][i][n][x]
elif kmer[j-1] == "C":
pwm[runnum][i][n][j]["C"] += hamdict[runnum][i][n][x]
elif kmer[j-1] == "T":
pwm[runnum][i][n][j]["T"] += hamdict[runnum][i][n][x]
elif kmer[j-1] == "G":
pwm[runnum][i][n][j]["G"] += hamdict[runnum][i][n][x]
for x in rhamdict[runnum][i][n]:
kmer = revComp(hash2kmer((x),i))
for j in range(1,i+1):
if kmer[j-1] == "A":
pwm[runnum][i][n][j]["A"] += rhamdict[runnum][i][n][x]
elif kmer[j-1] == "C":
pwm[runnum][i][n][j]["C"] += rhamdict[runnum][i][n][x]
elif kmer[j-1] == "T":
pwm[runnum][i][n][j]["T"] += rhamdict[runnum][i][n][x]
elif kmer[j-1] == "G":
pwm[runnum][i][n][j]["G"] += rhamdict[runnum][i][n][x]
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 listhammer():
for runnum in range(1, numofruns+1):
for i in kmercount[runnum]:
hamlist[runnum][i] = {}
rhamlist[runnum][i] = {}
for n in range(1, nmotifs+1):
hamlist[runnum][i][n] = []
rhamlist[runnum][i][n] = []
if n == 1:
hconsensus = max(kmercount[runnum][i], key=lambda key: kmercount[runnum][i][key])
else:
temptop6 = top6all[runnum][i].copy()
for x in temptop6:
if x in removelist[runnum][i][n]:
temptop6.remove(x)
temptop6.remove(kmer2hash(revComp(hash2kmer(x, i))))
hconsensus = max(temptop6, key=lambda key: kmercount[runnum][i][key])
consensus = hash2kmer(hconsensus, i)
for x in list(kmercount[runnum][i].keys()):
values = hash2kmer(x,i)
if hamming_distance(consensus, values) <= allowham:
if x not in removelist[runnum][i][n]:
rvalue = revComp(values)
rx = kmer2hash(rvalue)
try:
if kmercount[runnum][i][x] >= kmercount[runnum][i][rx]:
hamlist[runnum][i][n].append(x)
for j in range(n+1, nmotifs+1):
removelist[runnum][i][j].add(x)
removelist[runnum][i][j].add(rx)
if x != rx:
rhamlist[runnum][i][n].append(rx)
if n != nmotifs:
for j in range(n+1, nmotifs+1):
removelist[runnum][i][j].add(x)
removelist[runnum][i][j].add(rx)
except:
hamlist[runnum][i][n].append(x)
for j in range(n+1, nmotifs+1):
removelist[runnum][i][j].add(x)
removelist[runnum][i][j].add(rx)
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 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 top6(k):
top6s = []
keys = list(kmercount[(len(kmercount) - 1)][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 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 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()
