import matplotlib.pyplot as plt
import numpy as np
class RefDataType:
def __init__(self,length,steps,coverage,cv,marker,color,label):
self.length = length
self.steps = steps
self.coverage = coverage
self.cv = cv
self.marker = marker
self.color = color
self.label = label
def get_prop(self, prop_str):
if prop_str == 'length':
return self.length
elif prop_str == 'steps':
return self.steps
elif prop_str == 'coverage':
return self.coverage
elif prop_str == 'cv':
return self.cv
elif prop_str == 'marker':
return self.marker
elif prop_str == 'color':
return self.color
elif prop_str == 'label':
return self.label
else:
return "oh crap!"
def filter_prob_vals(vals, cnts):
""" replaces cases having no instances
with last value observed, notes where they occur
for plotting
"""
prob_vals = []
for ind, v in enumerate(vals):
if v == 0 and ind != 0 and cnts[ind]==0:
vals[ind] = vals[ind-1]
prob_vals.append(ind)
return vals, prob_vals
def ref_prop_plot(ref, prop, prop_ind, ranges):
if prop == 'steps':
x_vals = get_exp_x_mids(ranges[prop_ind],2)
elif prop == 'coverage':
x_vals = get_exp_x_mids(ranges[prop_ind],10)
else:
x_vals = get_x_mids(ranges[prop_ind])
num_meas = sum([b[1] for b in ref.get_prop(prop)])
# s=[a*1000./num_meas for a in [b[1] for b in ref.get_prop(prop)]]
y_vals = [b[2] for b in ref.get_prop(prop)]
y_vals, probs = filter_prob_vals(y_vals, [b[1] for b in ref.get_prop(prop)])
for i in range(len(y_vals)):
if i == len(y_vals)-1 and prop_ind==3:
axarr[0][prop_ind].scatter(x_vals[i], y_vals[i],
marker=ref.marker, label = ref.get_prop('label'),
facecolors=ref.color, edgecolors=ref.color, s=100)
elif i not in probs:
axarr[0][prop_ind].scatter(x_vals[i], y_vals[i],
marker=ref.marker,
facecolors=ref.color, edgecolors=ref.color, s=100)
else:
axarr[0][prop_ind].scatter(x_vals[i], y_vals[i],
marker=ref.marker,
facecolors='none', edgecolors=ref.color, s=1000)
axarr[0][prop_ind].plot(x_vals, y_vals, c=ref.color)
def ref_counts_plot(ref, prop, prop_ind, ranges):
x_vals = get_x_mids(ranges[prop_ind])
row = ref.get_prop(prop)
instances = np.array([a[1] for a in row])
instances = np.true_divide(instances, sum(instances))
axarr[1][prop_ind].scatter(x_vals, instances,
c=ref.color, marker=ref.marker, s=100)
axarr[1][prop_ind].plot(x_vals, instances, c=ref.color)
def get_x_mids(rng):
return 0.5 * ( np.array(rng[:-1]) + np.array(rng[1:]) )
def get_exp_x_mids(rng, base):
if base == 10:
vals = np.log10(rng)
else:
vals = np.log2(rng)
return base**get_x_mids(vals)
length_x_rng = [0,4000,8000,12000,16000,20000]
step_x_rng = [1,2,4,8,16,32]
cov_x_rng = [1,10,100,1e3,1e4,1e5]
cv_x_rng = [0, 0.05, 0.1, 0.15, 0.20, 0.25]
ranges = [length_x_rng, step_x_rng, cov_x_rng, cv_x_rng]
# input data - from alignment summary (makefile) output
# rows 0 = length, 1 = steps, 2 = coverage, 3 = CV
# plotting rows 4 = marker, 5 = color
ref_100 = RefDataType(
[(40, 46, 0.87), (19, 21, 0.9), (3, 6, 0.5), (4, 6, 0.67), (6, 7, 0.86)],
[(59, 59, 1.0), (6, 11, 0.55), (3, 10, 0.3), (3, 4, 0.75), (1, 2, 0.5)],
[(4, 4, 1.0), (39, 49, 0.8), (22, 26, 0.85), (7, 7, 1.0), (0, 0, 0)],
[(67, 73, 0.92), (4, 5, 0.8), (1, 4, 0.25), (0, 3, 0.0), (0, 1, 0.0)],
'^', 'y', '100'
)
ref_200 = RefDataType(
[(59, 73, 0.81), (35, 46, 0.76), (6, 11, 0.55), (3, 6, 0.5), (4, 6, 0.67)],
[(82, 83, 0.99), (3, 12, 0.25), (12, 24, 0.5), (7, 16, 0.44), (3, 7, 0.43)],
[(6, 8, 0.75), (55, 71, 0.77), (37, 52, 0.71), (8, 10, 0.8), (1, 1, 1.0)],
[(90, 95, 0.95), (7, 17, 0.41), (4, 11, 0.36), (3, 7, 0.43), (3, 11, 0.27)],
'v', 'g', '200'
)
ref_400 = RefDataType(
[(98, 118, 0.83), (62, 79, 0.78), (22, 27, 0.81), (13, 18, 0.72), (10, 12, 0.83)],
[(146, 147, 0.99), (24, 35, 0.69), (17, 39, 0.44), (11, 19, 0.58), (7, 13, 0.54)],
[(17, 22, 0.77), (105, 135, 0.78), (67, 77, 0.87), (11, 14, 0.79), (5, 6, 0.83)],
[(174, 188, 0.93), (13, 23, 0.57), (7, 14, 0.5), (8, 14, 0.57), (3, 15, 0.2)],
'h', 'c', '400'
)
ref_800 = RefDataType(
[(193, 236, 0.82), (107, 145, 0.74), (39, 55, 0.71), (18, 28, 0.64), (8, 13, 0.62)],
[(271, 278, 0.97), (39, 83, 0.47), (30, 58, 0.52), (10, 24, 0.42), (13, 27, 0.48)],
[(30, 46, 0.65), (174, 230, 0.76), (127, 162, 0.78), (21, 26, 0.81), (6, 6, 1.0)],
[(310, 345, 0.9), (25, 53, 0.47), (12, 27, 0.44), (7, 23, 0.3), (11, 27, 0.41)],
's', 'r', '800'
)
ref_1600 = RefDataType(
[(325, 404, 0.8), (181, 225, 0.8), (46, 72, 0.64), (35, 53, 0.66), (19, 25, 0.76)],
[(432, 442, 0.98), (72, 130, 0.55), (48, 95, 0.51), (29, 58, 0.5), (19, 40, 0.47)],
[(70, 104, 0.67), (253, 328, 0.77), (134, 173, 0.77), (119, 137, 0.87), (20, 23, 0.87)],
[(500, 548, 0.91), (46, 71, 0.65), (25, 50, 0.5), (23, 62, 0.37), (12, 48, 0.25)],
'o', 'b', '1600'
)
# plots
props = ['length', 'steps', 'coverage', 'cv']
f, axarr = plt.subplots(2, len(props), sharey='row', sharex='col')
# axarr[0].set_xticklabels(labels)
for ind, prop in enumerate(props):
for ref in [ref_100, ref_200, ref_400, ref_800, ref_1600]:
# print ref, prop, ind
ref_prop_plot(ref, prop, ind, ranges)
ref_counts_plot(ref, prop, ind, ranges)
if prop == 'cv':
axarr[0][ind].set_xlim(left=0, right=0.25)
axarr[1][ind].set_xlim(left=0, right=0.25)
if prop == 'steps':
axarr[0][ind].set_xscale('log', basex=2)
axarr[0][ind].set_xlim(1,32)
axarr[1][ind].set_xscale('log', basex=2)
axarr[1][ind].set_xlim(1,32)
if prop == 'coverage':
axarr[0][ind].set_xscale('log', basex=10)
axarr[1][ind].set_xscale('log', basex=10)
axarr[0][ind].set_title(prop.upper())
# x_vals = get_x_mids(ranges[ind])
# row = ref_1600.get_prop(prop)
# instances = np.array([a[1] for a in row])
# instances = np.true_divide(instances, sum(instances))
# axarr[1][ind].scatter(x_vals, instances,
# c=ref_1600.color, marker=ref_1600.marker, s=100)
# axarr[1][ind].plot(x_vals, instances, c=ref_1600.color)
# legend - put it on the rightmost
axarr[0][ind].legend()
plt.show()
