def hist_concordance(self, bins=100, fontsize=16):
"""
formula : 1 - (in + del + mismatch / (in + del + mismatch + match) )
For BWA and BLASR, the get_cigar_stats are different !!!
BWA for instance has no X stored while Pacbio forbids the use of the M
(CMATCH) tag. Instead, it uses X (CDIFF) and = (CEQUAL) characters.
Subread Accuracy: The post-mapping accuracy of the basecalls.
Formula: [1 - (errors/subread length)], where errors = number of deletions +
insertions + substitutions.
"""
try:
concordance = self._concordance
except:
self._set_concordance()
concordance = self._concordance
pylab.hist(concordance, bins=bins)
pylab.grid()
mu = np.mean(concordance)
median = np.median(concordance)
pylab.axvline(mu, color='r', alpha=0.5)
pylab.axvline(median, color='r', alpha=0.5, ls="--")
pylab.xlabel("concordance", fontsize=fontsize)
def hist_concordance(self, method, bins=100, fontsize=16):
"""
formula : 1 - (in + del + mismatch / (in + del + mismatch + match) )
For BWA and BLASR, the get_cigar_stats are different !!!
BWA for instance has no X stored while Pacbio forbids the use of the M
(CMATCH) tag. Instead, it uses X (CDIFF) and = (CEQUAL) characters.
Subread Accuracy: The post-mapping accuracy of the basecalls.
Formula: [1 - (errors/subread length)], where errors = number of deletions +
insertions + substitutions.
"""
try:
concordance = self._concordance
except:
self._set_concordance(method)
concordance = self._concordance
pylab.hist(concordance, bins=bins)
pylab.grid()
mu = np.mean(concordance)
median = np.median(concordance)
pylab.axvline(mu, color='r', alpha=0.5)
pylab.axvline(median, color='r', alpha=0.5, ls="--")
pylab.xlabel("concordance", fontsize=fontsize)
def get_stats(self, output="json"):
"""Return basic stats about the coverage data"""
data = self.df
stats = {
'DOC': self.df['cov'].mean(),
'STD': self.df['cov'].std(),
'Median': self.df['cov'].median(),
'BOC': 100 * sum(self.df['cov'] > 0) / float(len(self.df))}
try:
stats['CV'] = stats['STD'] / stats['DOC']
except:
stats['CV'] = np.nan
stats['MAD'] = np.median(abs(data['cov'].median() -
data['cov']).dropna())
names = ['BOC', 'CV', 'DOC', 'MAD', 'Median', 'STD']
descriptions = [
"breadth of coverage: the proportion (in %s) of the "
"genome covered by at least one read.",
"the coefficient of variation.",
"the sequencing depth (Depth of Coverage), that is the average of "
"the genome coverage.",
"median of the absolute median deviation defined as median(|X-median(X)|).",
"Median of the coverage.",
"standard deviation."
]
if 'gc' in self.df.columns:
stats['GC'] = self.df['gc'].mean() * 100
names.append('GC')
descriptions.append("GC content in %")
df = pd.DataFrame({
"name": names,
"Value": [stats[x] for x in names],
"Description": descriptions})
if output == "json":
return df.to_json()
else:
return df