def testSlidingZscore(self, ):
print "sliding zscore"
df = pd.DataFrame()
mean_abundance = dfc.calc_mean_abundance(self.elution, self.elution2)
mean_abundance.name = 'mean_abundance'
df = df.join(mean_abundance, how='outer')
feature_series = dfc.calc_diffrac(self.elution, self.elution2, normalize_totalCounts=True)
feature_series.name = 'diffrac_normalized'
df = df.join(feature_series, how='outer')
series = pd.Series([True,True,False,False,False],index= ['a','b','c','d','e'])
series.name = 'annotated'
df = df.join(series, how='outer')
print df
sliding_zscores = dfc.calc_sliding_zscore(df, window=1)
print sliding_zscores
np.testing.assert_almost_equal(sliding_zscores.loc['e']['sliding_zscore'],103.22222222222224)
sliding_zscores = dfc.calc_sliding_zscore(df, window=1, use_gmm=True)
print sliding_zscores
np.testing.assert_almost_equal(sliding_zscores.loc['e']['sliding_zscore'],103.22222222222224)
print "testing window = 5"
sliding_zscores = dfc.calc_sliding_zscore(df, window=5)
print sliding_zscores
np.testing.assert_almost_equal(sliding_zscores.loc['e']['sliding_zscore'],103.22222222222224)
print "testing gmm with artifically low min weight threshold 0.2"
sliding_zscores = dfc.calc_sliding_zscore(df, window=5, use_gmm=True, min_weight_threshold=0.2)
print sliding_zscores
#np.testing.assert_almost_equal(sliding_zscores.loc['e']['sliding_zscore'],529.68150807816676) #kdrew: had some instability because both components had equal weight
np.testing.assert_almost_equal(sliding_zscores.loc['b']['sliding_zscore'],-1.2864445660676826)
def testSlidingZscoreFDR(self, ):
print "sliding zscore FDR"
df = pd.DataFrame()
mean_abundance = dfc.calc_mean_abundance(self.elution, self.elution2)
mean_abundance.name = 'mean_abundance'
df = df.join(mean_abundance, how='outer')
feature_series = dfc.calc_diffrac(self.elution,
self.elution2,
normalize_totalCounts=True)
feature_series.name = 'diffrac_normalized'
df = df.join(feature_series, how='outer')
series = pd.Series([True, True, False, False, False],
index=['a', 'b', 'c', 'd', 'e'])
series.name = 'annotated'
df = df.join(series, how='outer')
sliding_zscores = dfc.calc_sliding_zscore(df, window=1)
df = df.join(sliding_zscores, how='outer')
fdr_pvals = dfc.calc_sliding_fdr_correct(df)
print fdr_pvals
np.testing.assert_almost_equal(fdr_pvals.loc['a']['sliding_pvalues'],
0.848460)
np.testing.assert_almost_equal(fdr_pvals.loc['e']['sliding_pvalues'],
0.0)
np.testing.assert_almost_equal(
fdr_pvals.loc['a']['sliding_pvalues_fdrcor'], 0.850628)
np.testing.assert_almost_equal(
fdr_pvals.loc['e']['sliding_pvalues_fdrcor'], 0.0)
def testSlidingZscore_nan(self, ):
print "sliding zscore2"
df = pd.DataFrame()
mean_abundance = dfc.calc_mean_abundance(self.elution3, self.elution4)
mean_abundance.name = 'mean_abundance'
df = df.join(mean_abundance, how='outer')
feature_series = dfc.calc_diffrac(self.elution3,
self.elution4,
normalize_totalCounts=True)
feature_series.name = 'diffrac_normalized'
df = df.join(feature_series, how='outer')
#series = pd.Series([True,True,False,False,False],index= ['a','b','c','d','e'])
#series.name = 'annotated'
#df = df.join(series, how='outer')
print df
sliding_zscores = dfc.calc_sliding_zscore(df, window=1)
print sliding_zscores
#kdrew: a is nan because there is no defined standard deviation within in the window, (b and c are both 1.0
np.testing.assert_almost_equal(
sliding_zscores.loc['a']['sliding_zscore'], np.nan)
def testDifFrac(self, ):
#kdrew: subtracts one dataframe from another and sums the absolute values (L1 norm)
diffrac_sum = dfc.calc_diffrac(self.elution, self.elution2)
print diffrac_sum
assert(diffrac_sum.loc['a'] == 0.0)
assert(diffrac_sum.loc['b'] == 1.0)
assert(diffrac_sum.loc['c'] == 2.0)
assert(diffrac_sum.loc['d'] == 4.0)
assert(diffrac_sum.loc['e'] == 17.0)
def testSlidingZscore(self, ):
print "sliding zscore"
df = pd.DataFrame()
mean_abundance = dfc.calc_mean_abundance(self.elution, self.elution2)
mean_abundance.name = 'mean_abundance'
df = df.join(mean_abundance, how='outer')
feature_series = dfc.calc_diffrac(self.elution,
self.elution2,
normalize_totalCounts=True)
feature_series.name = 'diffrac_normalized'
df = df.join(feature_series, how='outer')
series = pd.Series([True, True, False, False, False],
index=['a', 'b', 'c', 'd', 'e'])
series.name = 'annotated'
df = df.join(series, how='outer')
print df
sliding_zscores = dfc.calc_sliding_zscore(df, window=1)
print "Sliding_zscores"
print sliding_zscores
#kdrew: these scores changed alot after commit 17182d3, the crucial change was to the line that queries for gt_entries, mean_abundance went from > to >=
#kdrew: updated diffrac script to remove current index, now matches again
np.testing.assert_almost_equal(
sliding_zscores.loc['e']['sliding_zscore'], 103.22222222222224)
sliding_zscores = dfc.calc_sliding_zscore(df, window=1, use_gmm=True)
print sliding_zscores
np.testing.assert_almost_equal(
sliding_zscores.loc['e']['sliding_zscore'], 103.22222222222224)
print "testing window = 5"
sliding_zscores = dfc.calc_sliding_zscore(df, window=5)
print sliding_zscores
np.testing.assert_almost_equal(
sliding_zscores.loc['e']['sliding_zscore'], 103.22222222222224)
print "testing gmm with artifically low min weight threshold 0.2"
sliding_zscores = dfc.calc_sliding_zscore(df,
window=5,
use_gmm=True,
min_weight_threshold=0.2)
print sliding_zscores
#np.testing.assert_almost_equal(sliding_zscores.loc['e']['sliding_zscore'],529.68150807816676) #kdrew: had some instability because both components had equal weight
np.testing.assert_almost_equal(
sliding_zscores.loc['b']['sliding_zscore'], -1.2864445660676826)
def testZscore(self, ):
print "zscore"
df = pd.DataFrame()
mean_abundance = dfc.calc_mean_abundance(self.elution, self.elution2)
mean_abundance.name = 'mean_abundance'
df = df.join(mean_abundance, how='outer')
feature_series = dfc.calc_diffrac(self.elution, self.elution2, normalize_totalCounts=True)
feature_series.name = 'diffrac_normalized'
df = df.join(feature_series, how='outer')
series = pd.Series([True,True,False,False,False],index= ['a','b','c','d','e'])
series.name = 'annotated'
df = df.join(series, how='outer')
print df
zscores = dfc.calc_zscore(df)
print zscores.loc['e']
np.testing.assert_almost_equal(zscores.loc['e'],1.4140145096382672)