def test_composition():
"Test sequence compositions"
# load data
df = txt_to_df(path + '/tests/docs/test_seq.txt', 0)
# test relative composition
comp_rel = composition(df, 'relative')
# test if frequencies add to 1
for i in range(df.shape[0]):
assert round(comp_rel.iloc[0, :].sum()) == 1
# test decimal places
comp_dec2 = composition(df, 'relative', round_fraction=2)
x = str(comp_dec2['A'][0])
assert x[::-1].find('.') == 2
comp_dec2 = composition(df, 'relative', round_fraction=5)
y = str(comp_dec2['A'][0])
assert y[::-1].find('.') == 5
# test absolute composition
comp_abs = composition(df, 'absolute')
# test if frequences == sequence length
all_lengths = [6, 9, 7, 6]
for i in range(df.shape[0]):
assert comp_abs.iloc[i, :].sum() == all_lengths[i]
def test_aaindex3():
"Test AAIndex3"
# load data
df = txt_to_df(path + '/tests/docs/test_seq.txt', 0)
# get aaindex2
aaind3 = aaindex3(df)
# test shape
assert aaind3.shape == (4, 43)
# test some triangular indices
TANS760101 = np.array([-4.72, -5.975, -4.18333, -4.04])
GODA950101 = np.array([np.nan, -.05, -.1333, -.14])
ZHAC000106 = np.array([.196, -.34875, .46666, .972])
np.testing.assert_equal(np.round(aaind3['TANS760101'], 3),\
np.round(TANS760101, 3))
# this column contains NaNs
assert ('GODA950101' in aaind3) == False
np.testing.assert_equal(np.round(aaind3['ZHAC000106'], 3),\
np.round(ZHAC000106, 3))
# test some square indices
ZHAC000102 = np.array([-.408, -1.415, .475, 1.532])
ZHAC000103 = np.array([-.052, -.625, .59166, 1.096])
ZHAC000105 = np.array([-.242, -.72, .17, .952])
np.testing.assert_equal(np.round(aaind3['ZHAC000102'], 3),\
np.round(ZHAC000102, 3))
np.testing.assert_equal(np.round(aaind3['ZHAC000103'], 3),\
np.round(ZHAC000103, 3))
np.testing.assert_equal(np.round(aaind3['ZHAC000105'], 3),\
np.round(ZHAC000105, 3))
# test standardization (zscore)
aaind3_z = aaindex3(df, 'zscore')
# test mean = 0
for i in range(aaind3_z.shape[0]):
assert abs(round(aaind3_z.iloc[:, 1].mean())) == 0
# test std --> 1
for i in range(aaind3_z.shape[0]):
assert round(aaind3_z.iloc[:,i].std(), 1) ==\
round(aaind3_z.iloc[:,0].std(), 1)
# test standardization (minmax)
aaind3_mm = aaindex3(df, 'minmax')
# test minimum and maximum
for i in range(aaind3_mm.shape[0]):
assert round(aaind3_mm.iloc[:, i].min()) == 0
assert round(aaind3_mm.iloc[:, i].max()) == 1
def test_aaindex2():
"Test AAIndex2"
# load data
df = txt_to_df(path + '/tests/docs/test_seq.txt', 0)
# get aaindex2
aaind2 = aaindex2(df)
# test shape
assert aaind2.shape == (4, 94)
# test some triangular indices
ALTS910101 = np.array([-2, -.125, .333, -2])
VOGG950101 = np.array([4.28, 5.2, 6.05, 4.32])
CROG050101 = np.array([-1.8, .625, .5, -.4])
np.testing.assert_equal(np.round(aaind2['ALTS910101'], 3),\
np.round(ALTS910101, 3))
np.testing.assert_equal(np.round(aaind2['VOGG950101'], 3),\
np.round(VOGG950101, 3))
np.testing.assert_equal(np.round(aaind2['CROG050101'], 3),\
np.round(CROG050101, 3))
# test some square indices
LINK010101 = np.array([.0266, .0955, .13, .1276])
KOSJ950108 = np.array([1.62, 18.0875, 16.05, 14.68])
DOSZ010101 = np.array([1.32, 15.7625, -1.1833, -5.12])
np.testing.assert_equal(np.round(aaind2['LINK010101'], 3),\
np.round(LINK010101, 3))
np.testing.assert_equal(np.round(aaind2['KOSJ950108'], 3),\
np.round(KOSJ950108, 3))
np.testing.assert_equal(np.round(aaind2['DOSZ010101'], 3),\
np.round(DOSZ010101, 3))
# test standardization (zscore)
aaind2_z = aaindex2(df, 'zscore')
# test mean = 0
for i in range(aaind2_z.shape[0]):
assert abs(round(aaind2_z.iloc[:, 1].mean())) == 0
# test std --> 1
for i in range(aaind2_z.shape[0]):
assert round(aaind2_z.iloc[:,i].std(), 1) ==\
round(aaind2_z.iloc[:,0].std(), 1)
# test standardization (minmax)
aaind2_mm = aaindex2(df, 'minmax')
# test minimum and maximum
for i in range(aaind2_mm.shape[0]):
assert round(aaind2_mm.iloc[:, i].min()) == 0
assert round(aaind2_mm.iloc[:, i].max()) == 1
def test_lengths():
"Test sequence lengths"
# load data
df = txt_to_df(path + '/tests/docs/test_seq.txt', 0)
# test integer lengths
len_int = length(df, 'int')
assert np.array_equal(len_int, np.array([6, 9, 7, 6]))
# test one-hot-encoded lengths
len_ohe = length(df, 'ohe')
# columns: [6, 7, 9]
assert np.array_equal(
len_ohe,
np.array([[1., 0., 0.], [0., 0., 1.], [0., 1., 0.], [1., 0., 0.]]))
def test_conversion():
"Test txt_to_df conversion"
# load data
df = txt_to_df(path+'/tests/docs/test_seq.txt', 0)
# test labels and df shape
assert df.columns[0] == 'Sequence'
assert df.columns[1] == 'Label'
assert df.shape == (4, 2)
# test sequences
assert df['Sequence'][0] == 'AGTYLK'
assert df['Sequence'][1] == 'VCIMMMPFP'
assert df['Sequence'][2] == 'LRSAHHN'
assert df['Sequence'][3] == 'AQEEWD'
# test labels
for i in range(df.shape[0]):
assert df['Label'][i] == 0
def test_aaindex1():
"Test AAIndex1"
# load data
df = txt_to_df(path + '/tests/docs/test_seq.txt', 0)
# get aaindex1
aaind1 = aaindex1(df)
# test shape
assert aaind1.shape == (4, 553)
# test some indices
ANDN920101 = np.array([4.3, 4.40555, 4.48714, 4.46])
QIAN880126 = np.array([.01166, -.17111, .05857, -.04333])
KARS160122 = np.array([2.014, 5.48522, 2.789, 1.751])
np.testing.assert_equal(np.round(aaind1['ANDN920101'], 3),\
np.round(ANDN920101, 3))
np.testing.assert_equal(np.round(aaind1['QIAN880126'], 3),\
np.round(QIAN880126, 3))
np.testing.assert_equal(np.round(aaind1['KARS160122'], 3),\
np.round(KARS160122, 3))
# test standardization (zscore)
aaind1_z = aaindex1(df, 'zscore')
# test mean = 0
for i in range(aaind1_z.shape[0]):
assert abs(round(aaind1_z.iloc[:, 1].mean())) == 0
# test std --> 1
for i in range(aaind1_z.shape[0]):
assert round(aaind1_z.iloc[:,i].std(), 1) ==\
round(aaind1_z.iloc[:,0].std(), 1)
# test standardization (minmax)
aaind1_mm = aaindex1(df, 'minmax')
# test minimum and maximum
for i in range(aaind1_mm.shape[0]):
assert round(aaind1_mm.iloc[:, i].min()) == 0
assert round(aaind1_mm.iloc[:, i].max()) == 1
def test_integer_encode():
"Test integer encoding"
# load data
df = txt_to_df(path + '/tests/docs/test_seq.txt', 0)
enc = integer_encode(df)
# test array shape and type
assert enc.shape == (4, )
assert type(enc) == np.ndarray
# test array contents
assert np.array_equal(enc[0], np.array([1, 6, 17, 20, 10, 9]))
assert np.array_equal(enc[1], np.array([18, 2, 8, 11, 11, 11, 13, 5, 13]))
assert np.array_equal(enc[2], np.array([10, 15, 16, 1, 7, 7, 12]))
assert np.array_equal(enc[3], np.array([1, 14, 4, 4, 19, 3]))
# test padding
enc = integer_encode(df, padding=True)
assert enc.shape == (4, 9)
assert [enc[0][i] == 0 for i in [6, 7, 8]]
assert [enc[2][i] == 0 for i in [7, 8]]
assert [enc[3][i] == 0 for i in [6, 7, 8]]