def test_region_check(self):
with self.assertRaises(genome_tools.BadRegionError):
self.genome.check_region(genome_tools.Region('chr4',10,20))
with self.assertRaises(genome_tools.BadRegionError):
self.genome.check_region(genome_tools.Region('chr1',-1, 20))
with self.assertRaises(genome_tools.BadRegionError):
self.genome.check_region(genome_tools.Region('chr2', 300, 400))
def test_get_window_from_position(self):
region, window_idx = self.genome.get_window_from_position('chr2', 250)
self.assertTrue(
region == genome_tools.Region('chr2', 200, 300)
)
next_region, next_window_idx = self.genome.get_next_window(region)
self.assertTrue(
next_region == genome_tools.Region('chr2',300,375)
)
self.assertTrue(
window_idx == (next_window_idx - 1)
)
def _check_region_specification(self, regions):
invalid_chroms = Counter()
valid_regions = []
for i, region in enumerate(regions):
assert(isinstance(region, (tuple, list)) and len(region) == 3), 'Error at region #{}: Each region passed must be in format (string \"chr\",int start, int end'\
.format(str(i))
try:
new_region = genome_tools.Region(*region, annotation=i)
self.data_interface.genome.check_region(new_region)
valid_regions.append(new_region)
except ValueError:
raise AssertionError(
'Error at region #{}: Could not coerce positions into integers'
.format(str(i)))
except genome_tools.NotInGenomeError as err:
invalid_chroms[region[0]] += 1
#raise AssertionError('Error at region #{}: '.format(str(i+1)) + str(err) + '\nOnly main chromosomes (chr[1-22,X,Y] for hg38, and chr[1-19,X,Y] for mm10) are permissible for LISA test.')
except genome_tools.BadRegionError as err:
raise AssertionError('Error at region #{}: '.format(str(i +
1)) +
str(err))
if len(invalid_chroms) > 0:
self.log.append(
'WARNING: {} regions encounted from unknown chromsomes: {}'.
format(str(sum(invalid_chroms.values())),
str(','.join(invalid_chroms.keys()))))
return valid_regions
def test_get_window_idx(self):
self.assertTrue(
self.genome.get_region(0)[0] == genome_tools.Region('chr1', 0, 100)
)
self.assertTrue(
self.genome.get_region(4)[0] == genome_tools.Region('chr1',400,450)
)
self.assertTrue(
self.genome.get_region(5)[0] == genome_tools.Region('chr2',0,100)
)
self.assertTrue(
self.genome.get_region(14)[0] == genome_tools.Region('chr3',500,600)
)
with self.assertRaises(AssertionError):
self.genome.get_region(15)
def test_overlap_any(self):
self.assertTrue(
genome_tools.Region('chr1',5,10).overlaps(genome_tools.Region('chr1',8,15))
)
self.assertFalse(
genome_tools.Region('chr1',1,11).overlaps(genome_tools.Region('chr1',9,15), min_overlap_proportion=0.4)
)
self.assertTrue(
genome_tools.Region('chr1',1,11).overlaps(genome_tools.Region('chr1',6,15), min_overlap_proportion=0.4)
)
def test_invalid_region(self):
with self.assertRaises(AssertionError):
genome_tools.Region('chr1',10, 5)
def test_regions_equal(self):
self.assertTrue(
genome_tools.Region('chr1',10,15) == genome_tools.Region('chr1', 10, 15)
)
def test_distance_function(self):
self.assertEqual(
genome_tools.Region('chr1',20, 30).get_genomic_distance(genome_tools.Region('chr1', 120, 130)), 100)
def test_overlap_negative_abutted(self):
self.assertFalse(
genome_tools.Region('chr1',5,10).overlaps(genome_tools.Region('chr1',10,20))
)
def test_overlap_negative(self):
self.assertFalse(
genome_tools.Region('chr1',5,10).overlaps(genome_tools.Region('chr1',20,25))
)
def setUp(self):
self.genome = genome_tools.Genome(['chr1','chr2','chr3'], [450, 375, 600], window_size=50)
self.scrambled_genome = genome_tools.Genome(['chr1','chr3','chr2'], [450, 600, 375], window_size=50)
self.regions1A = [
genome_tools.Region('chr1',20,40),
genome_tools.Region('chr1',30,60),
genome_tools.Region('chr1',210,230),
genome_tools.Region('chr2',100,150),
genome_tools.Region('chr2',220,233),
genome_tools.Region('chr3',430,450),
]
self.regions1B = [
genome_tools.Region('chr1',20,40),
genome_tools.Region('chr1',30,60),
genome_tools.Region('chr1',210,230),
genome_tools.Region('chr2',100,150),
genome_tools.Region('chr2',220,233),
genome_tools.Region('chr3',430,450),
]
self.auto_distancing_truth = np.array(
[[0,15,0,0,0,0],
[15,0,175,0,0,0],
[0,175,0,0,0,0],
[0,0,0,0,101,0],
[0,0,0,101,0,0],
[0,0,0,0,0,0]
]
)
self.m2m_map_truth = np.array([
(0,0),
(0,1),
(1,1),
(4,2),
(11,3),
(13,4),
(25,5)
])
def test_overlap_different_chr(self):
self.assertFalse(
genome_tools.Region('chr1',5,10).overlaps(genome_tools.Region('chr2',5,10))
)