def test_NNC_match(self):
""" Example where the transcript is an NNC match to an existing one by
virtue of a new splice donor.
"""
conn, cursor = get_db_cursor()
build = "toy_build"
database = "scratch/toy.db"
talon.get_counters(database)
edge_dict = init_refs.make_edge_dict(cursor)
location_dict = init_refs.make_location_dict(build, cursor)
run_info = talon.init_run_info(database, build)
transcript_dict = init_refs.make_transcript_dict(cursor, build)
vertex_2_gene = init_refs.make_vertex_2_gene_dict(cursor)
gene_starts = init_refs.make_gene_start_or_end_dict(
cursor, build, "start")
gene_ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
chrom = "chr1"
positions = [1, 110, 900, 1000]
edge_IDs = [talon.edge_counter.value() + 1]
vertex_IDs = [talon.vertex_counter.value() + 1, 5]
strand = "+"
v_novelty = [0, 0]
gene_ID, transcript_ID, transcript_novelty, start_end_info = talon.process_NNC(
chrom, positions, strand, edge_IDs, vertex_IDs, transcript_dict,
gene_starts, gene_ends, edge_dict, location_dict, vertex_2_gene,
run_info)
correct_gene_ID = fetch_correct_ID("TG1", "gene", cursor)
assert gene_ID == correct_gene_ID
assert start_end_info["vertex_IDs"] == [1] + vertex_IDs + [6]
assert transcript_dict[frozenset(start_end_info["edge_IDs"])] != None
conn.close()
def test_no_match(self):
""" Example with no FSM match """
conn, cursor = get_db_cursor()
build = "toy_build"
db = "scratch/toy.db"
talon.get_counters(db)
edge_dict = init_refs.make_edge_dict(cursor)
location_dict = init_refs.make_location_dict(build, cursor)
run_info = talon.init_run_info(db, build)
transcript_dict = init_refs.make_transcript_dict(cursor, build)
gene_starts = init_refs.make_gene_start_or_end_dict(
cursor, build, "start")
gene_ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
chrom = "chr1"
positions = [1, 100, 500, 600]
strand = "+"
edge_IDs = [2]
vertex_IDs = [2, 3, 4, 5]
v_novelty = [0, 0, 0, 0]
all_matches = talon.search_for_ISM(edge_IDs, transcript_dict)
gene_ID, transcript_ID, novelty, start_end_info = talon.process_FSM(
chrom, positions, strand, edge_IDs, vertex_IDs, all_matches,
gene_starts, gene_ends, edge_dict, location_dict, run_info)
assert gene_ID == transcript_ID == None
conn.close()
def test_no_match(self):
""" Example with no ISM match """
conn, cursor = get_db_cursor()
build = "toy_build"
database = "scratch/toy.db"
run_info = talon.init_run_info(database, build)
talon.get_counters(database)
edge_dict = init_refs.make_edge_dict(cursor)
location_dict = init_refs.make_location_dict(build, cursor)
transcript_dict = init_refs.make_transcript_dict(cursor, build)
gene_starts = init_refs.make_gene_start_or_end_dict(cursor, build, "start")
gene_ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
chrom = "chr1"
strand = "+"
positions = [ 1, 100, 900, 1000 ]
edge_IDs = [200]
vertex_IDs = [2, 5]
v_novelty = [0, 0]
all_matches = talon.search_for_ISM(edge_IDs, transcript_dict)
assert all_matches == None
conn.close()
def test_match(self):
""" Example where the transcript is a monoexonic match.
"""
conn, cursor = get_db_cursor()
build = "toy_build"
database = "scratch/toy.db"
talon.get_counters(database)
init_refs.make_temp_novel_gene_table(cursor, build)
init_refs.make_temp_monoexonic_transcript_table(cursor, build)
edge_dict = init_refs.make_edge_dict(cursor)
location_dict = init_refs.make_location_dict(build, cursor)
run_info = talon.init_run_info(database, build)
transcript_dict = init_refs.make_transcript_dict(cursor, build)
vertex_2_gene = init_refs.make_vertex_2_gene_dict(cursor)
gene_starts = init_refs.make_gene_start_or_end_dict(cursor, build, "start")
gene_ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
chrom = "chr4"
strand = "-"
positions = ( 3900, 1100 )
annotation = talon.identify_monoexon_transcript(chrom, positions,
strand, cursor,
location_dict, edge_dict,
transcript_dict, vertex_2_gene,
gene_starts, gene_ends, run_info,
'temp_gene', 'temp_monoexon')
correct_gene_ID = fetch_correct_ID("TG6", "gene", cursor)
correct_transcript_ID = fetch_correct_ID("TG6-001", "transcript", cursor)
assert annotation['gene_ID'] == correct_gene_ID
assert annotation['start_delta'] == 100
assert annotation['end_delta'] == -100
conn.close()
def test_all_genes(self):
""" Get starts and ends for all known genes in the database """
conn, cursor = get_db_cursor()
build = "toy_build"
starts = init_refs.make_gene_start_or_end_dict(cursor, build, "start")
ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
conn.close()
# Check starts. Field 1: Gene ID, Field 2: start pos, Field 3: vertex ID
assert starts == {1: {1: 1},
2: {2000: 8},
3: {5000: 9},
4: {1: 13},
5: {1: 19, 800: 23},
6: {4000: 34}}
assert ends == {1: {1000: 6},
2: {900: 5},
3: {6500: 12},
4: {1000: 18},
5: {2200: 30, 2600: 32},
6: {1000: 33}}
def test_antisense(self):
""" Example where the vertices are known but there is no same-strand
match """
conn, cursor = get_db_cursor()
build = "toy_build"
database = "scratch/toy.db"
talon.get_counters(database)
edge_dict = init_refs.make_edge_dict(cursor)
locations = init_refs.make_location_dict(build, cursor)
run_info = talon.init_run_info(database, build)
transcript_dict = init_refs.make_transcript_dict(cursor, build)
vertex_2_gene = init_refs.make_vertex_2_gene_dict(cursor)
gene_starts = init_refs.make_gene_start_or_end_dict(
cursor, build, "start")
gene_ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
# Construct temp novel gene db
init_refs.make_temp_novel_gene_table(cursor, "toy_build")
chrom = "chr1"
start = 1000
end = 1
edge_IDs = [talon.edge_counter.value() + 1]
positions = [1000, 900, 100, 1]
vertex_IDs = [5, 2]
strand = "-"
anti_strand = "+"
v_novelty = (0, 0, 0, 0)
# Find antisense match
gene_ID, transcript_ID, gene_novelty, transcript_novelty, start_end_info = \
talon.process_spliced_antisense(chrom, positions,
strand, edge_IDs,
vertex_IDs,
transcript_dict,
gene_starts,
gene_ends,
edge_dict, locations,
vertex_2_gene, run_info,
cursor, "temp_gene")
#anti_gene_ID = talon.find_gene_match_on_vertex_basis(vertex_IDs,
# anti_strand,
# vertex_2_gene)
correct_gene_ID = fetch_correct_ID("TG1", "gene", cursor)
anti_gene_ID = gene_novelty[-1][-1]
assert anti_gene_ID == correct_gene_ID
assert start_end_info["vertex_IDs"] == [6, 5, 2, 1]
conn.close()
def test_intergenic(self):
""" Example where the transcript is an NIC match to an existing one by
virtue of a new splice donor.
"""
conn, cursor = get_db_cursor()
build = "toy_build"
database = "scratch/toy.db"
talon.get_counters(database)
edge_dict = init_refs.make_edge_dict(cursor)
location_dict = init_refs.make_location_dict(build, cursor)
run_info = talon.init_run_info(database, build)
transcript_dict = init_refs.make_transcript_dict(cursor, build)
vertex_2_gene = init_refs.make_vertex_2_gene_dict(cursor)
gene_starts = init_refs.make_gene_start_or_end_dict(
cursor, build, "start")
gene_ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
correct_gene_ID = talon.gene_counter.value() + 1
# Construct temp novel gene db
init_refs.make_temp_novel_gene_table(cursor, "toy_build")
chrom = "chrX"
positions = [1, 100, 900, 1000]
edge_IDs = [
talon.edge_counter.value() + 1,
talon.edge_counter.value() + 2
]
vertex_IDs = [
talon.vertex_counter.value() + 1,
talon.vertex_counter.value() + 2
]
strand = "+"
gene_ID, transcript_ID, gene_novelty, transcript_novelty, start_end_info = \
talon.process_remaining_mult_cases(chrom, positions,
strand, edge_IDs,
vertex_IDs,
transcript_dict,
gene_starts, gene_ends,
edge_dict, location_dict,
vertex_2_gene, run_info,
cursor, "temp_gene")
assert gene_ID == correct_gene_ID
assert transcript_dict[frozenset(start_end_info["edge_IDs"])] != None
assert gene_novelty[0][-2] == "intergenic_novel"
conn.close()
def test_genomic(self):
""" Example where the transcript overlaps a gene but contains no known
splice vertices
"""
conn, cursor = get_db_cursor()
build = "toy_build"
database = "scratch/toy.db"
talon.get_counters(database)
edge_dict = init_refs.make_edge_dict(cursor)
location_dict = init_refs.make_location_dict(build, cursor)
run_info = talon.init_run_info(database, build)
transcript_dict = init_refs.make_transcript_dict(cursor, build)
vertex_2_gene = init_refs.make_vertex_2_gene_dict(cursor)
gene_starts = init_refs.make_gene_start_or_end_dict(
cursor, build, "start")
gene_ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
# Construct temp novel gene db
init_refs.make_temp_novel_gene_table(cursor, "toy_build")
chrom = "chr1"
positions = [1000, 950, 700, 600]
edge_IDs = [
talon.edge_counter.value() + 1,
talon.edge_counter.value() + 2
]
vertex_IDs = [
talon.vertex_counter.value() + 1,
talon.vertex_counter.value() + 2
]
strand = "-"
gene_ID, transcript_ID, gene_novelty, transcript_novelty, start_end_info = \
talon.process_remaining_mult_cases(chrom, positions,
strand, edge_IDs,
vertex_IDs,
transcript_dict,
gene_starts, gene_ends,
edge_dict, location_dict,
vertex_2_gene, run_info,
cursor, "temp_gene")
correct_gene_ID = fetch_correct_ID("TG3", "gene", cursor)
assert gene_ID == correct_gene_ID
assert transcript_dict[frozenset(start_end_info["edge_IDs"])] != None
assert gene_novelty == []
assert transcript_novelty[-1][-2] == "genomic_transcript"
conn.close()
def test_ISM_suffix(self):
""" Example where the transcript is an ISM with suffix
"""
conn, cursor = get_db_cursor()
build = "toy_build"
database = "scratch/toy.db"
run_info = talon.init_run_info(database, build)
talon.get_counters(database)
edge_dict = init_refs.make_edge_dict(cursor)
location_dict = init_refs.make_location_dict(build, cursor)
transcript_dict = init_refs.make_transcript_dict(cursor, build)
gene_starts = init_refs.make_gene_start_or_end_dict(cursor, build, "start")
gene_ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
chrom = "chr1"
strand = "+"
positions = [ 500, 600, 900, 1000 ]
edge_IDs = [4]
vertex_IDs = [4, 5]
v_novelty = [0, 0]
all_matches = talon.search_for_ISM(edge_IDs, transcript_dict)
gene_ID, transcript_ID, novelty, start_end_info = talon.process_ISM(chrom,
positions,
strand, edge_IDs,
vertex_IDs,
all_matches,
transcript_dict,
gene_starts, gene_ends,
edge_dict, location_dict,
run_info)
correct_gene_ID = fetch_correct_ID("TG1", "gene", cursor)
assert gene_ID == correct_gene_ID
assert start_end_info["vertex_IDs"] == [3, 4, 5, 6]
assert start_end_info["edge_IDs"] == [3, 4, 5]
assert start_end_info["start_novelty"] == 0 # because the exon is known
assert start_end_info["end_novelty"] == 0
assert transcript_dict[frozenset(start_end_info["edge_IDs"])] != None
conn.close()
def test_partial_match(self):
""" Example where the transcript overlaps a single-exon transcript,
but is shorter. In the past, the start would be assigned to the
annotated start, and the end would be novel. This is no longer
the case- at this time, the transcript will be assigned to
the annotated match. """
conn, cursor = get_db_cursor()
build = "toy_build"
database = "scratch/toy.db"
talon.get_counters(database)
init_refs.make_temp_novel_gene_table(cursor, build)
init_refs.make_temp_monoexonic_transcript_table(cursor, build)
edge_dict = init_refs.make_edge_dict(cursor)
location_dict = init_refs.make_location_dict(build, cursor)
run_info = talon.init_run_info(database, build)
transcript_dict = init_refs.make_transcript_dict(cursor, build)
vertex_2_gene = init_refs.make_vertex_2_gene_dict(cursor)
gene_starts = init_refs.make_gene_start_or_end_dict(cursor, build, "start")
gene_ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
chrom = "chr4"
strand = "-"
positions = ( 3900, 2900 )
annotation = talon.identify_monoexon_transcript(chrom, positions,
strand, cursor,
location_dict, edge_dict,
transcript_dict, vertex_2_gene,
gene_starts, gene_ends, run_info,
'temp_gene', 'temp_monoexon')
correct_gene_ID = fetch_correct_ID("TG6", "gene", cursor)
correct_transcript_ID = fetch_correct_ID("TG6-001", "transcript", cursor)
assert annotation['gene_ID'] == correct_gene_ID
assert annotation['transcript_ID'] == correct_transcript_ID
assert annotation['start_delta'] == 100
assert annotation['end_delta'] == -1900
conn.close()
def test_FSM_start_diff(self):
""" Example where the transcript is an FSM but has a difference on
the start large enough to be novel.
"""
conn, cursor = get_db_cursor()
build = "toy_build"
db = "scratch/toy.db"
talon.get_counters(db)
edge_dict = init_refs.make_edge_dict(cursor)
location_dict = init_refs.make_location_dict(build, cursor)
run_info = talon.init_run_info(db, build)
transcript_dict = init_refs.make_transcript_dict(cursor, build)
orig_vertices = talon.vertex_counter.value()
gene_starts = init_refs.make_gene_start_or_end_dict(
cursor, build, "start")
gene_ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
chrom = "chr1"
positions = [2501, 1500, 1000, 900] #First postion is > 500bp away
strand = "-"
edge_IDs = [7]
vertex_IDs = [7, 6]
v_novelty = [0, 0]
all_matches = talon.search_for_ISM(edge_IDs, transcript_dict)
gene_ID, transcript_ID, novelty, start_end_info = talon.process_FSM(
chrom, positions, strand, edge_IDs, vertex_IDs, all_matches,
gene_starts, gene_ends, edge_dict, location_dict, run_info)
correct_gene_ID = fetch_correct_ID("TG3", "gene", cursor)
correct_transcript_ID = fetch_correct_ID("TG3-001", "transcript",
cursor)
assert gene_ID == correct_gene_ID
assert transcript_ID == correct_transcript_ID
assert start_end_info["start_vertex"] == orig_vertices + 1
assert start_end_info["end_vertex"] == 5
conn.close()
def test_interval_1_2000(self):
""" Get starts and ends for genes in the database that overlap the
interval chr1:1-2000 """
conn, cursor = get_db_cursor()
build = "toy_build"
starts = init_refs.make_gene_start_or_end_dict(cursor, build, "start",
chrom = "chr1",
start = 1,
end = 2000)
ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end",
chrom = "chr1",
start = 1,
end = 2000)
conn.close()
assert starts == {1: {1: 1},
2: {2000: 8}}
assert ends == {1: {1000: 6},
2: {900: 5}}
def test_FSM_perfect(self):
""" Example where the transcript is a perfect full splice match.
"""
conn, cursor = get_db_cursor()
build = "toy_build"
db = "scratch/toy.db"
talon.get_counters(db)
edge_dict = init_refs.make_edge_dict(cursor)
location_dict = init_refs.make_location_dict(build, cursor)
run_info = talon.init_run_info(db, build)
transcript_dict = init_refs.make_transcript_dict(cursor, build)
gene_starts = init_refs.make_gene_start_or_end_dict(
cursor, build, "start")
gene_ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
chrom = "chr1"
positions = [1, 100, 500, 600, 900, 1010]
strand = "+"
edge_IDs = [2, 3, 4]
vertex_IDs = [2, 3, 4, 5]
v_novelty = [0, 0, 0, 0]
all_matches = talon.search_for_ISM(edge_IDs, transcript_dict)
gene_ID, transcript_ID, novelty, start_end_info = talon.process_FSM(
chrom, positions, strand, edge_IDs, vertex_IDs, all_matches,
gene_starts, gene_ends, edge_dict, location_dict, run_info)
correct_gene_ID = fetch_correct_ID("TG1", "gene", cursor)
correct_transcript_ID = fetch_correct_ID("TG1-001", "transcript",
cursor)
assert gene_ID == correct_gene_ID
assert transcript_ID == correct_transcript_ID
assert novelty == []
assert start_end_info["start_vertex"] == 1
assert start_end_info["end_vertex"] == 6
assert start_end_info["diff_3p"] == 10
conn.close()
def test_ISM_prefix(self):
""" Example where the transcript is a prefix ISM with a novel start
"""
conn, cursor = get_db_cursor()
build = "toy_build"
database = "scratch/toy.db"
run_info = talon.init_run_info(database, build)
talon.get_counters(database)
edge_dict = init_refs.make_edge_dict(cursor)
location_dict = init_refs.make_location_dict(build, cursor)
transcript_dict = init_refs.make_transcript_dict(cursor, build)
gene_starts = init_refs.make_gene_start_or_end_dict(cursor, build, "start")
gene_ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
chrom = "chr1"
strand = "+"
positions = [ 1, 100, 500, 600 ]
edge_IDs = [2]
vertex_IDs = [2, 3]
v_novelty = [0, 0]
all_matches = talon.search_for_ISM(edge_IDs, transcript_dict)
gene_ID, transcript_ID, novelty, start_end_info = talon.process_ISM(chrom,
positions,
strand, edge_IDs,
vertex_IDs,
all_matches,
transcript_dict,
gene_starts, gene_ends,
edge_dict, location_dict,
run_info)
correct_gene_ID = fetch_correct_ID("TG1", "gene", cursor)
assert gene_ID == correct_gene_ID
assert start_end_info["vertex_IDs"] == [1, 2, 3, 4]
assert start_end_info["edge_IDs"] == [1, 2, 3]
conn.close()
def test_antisense(self):
""" Example where the transcript is antisense """
conn, cursor = get_db_cursor()
build = "toy_build"
database = "scratch/toy.db"
talon.get_counters(database)
init_refs.make_temp_novel_gene_table(cursor, build)
init_refs.make_temp_monoexonic_transcript_table(cursor, build)
edge_dict = init_refs.make_edge_dict(cursor)
location_dict = init_refs.make_location_dict(build, cursor)
run_info = talon.init_run_info(database, build)
transcript_dict = init_refs.make_transcript_dict(cursor, build)
vertex_2_gene = init_refs.make_vertex_2_gene_dict(cursor)
gene_starts = init_refs.make_gene_start_or_end_dict(cursor, build, "start")
gene_ends = init_refs.make_gene_start_or_end_dict(cursor, build, "end")
chrom = "chr4"
strand = "+"
positions = ( 1300, 3900 )
annotation = talon.identify_monoexon_transcript(chrom, positions,
strand, cursor,
location_dict, edge_dict,
transcript_dict, vertex_2_gene,
gene_starts, gene_ends, run_info,
'temp_gene', 'temp_monoexon')
anti_gene_ID = fetch_correct_ID("TG6", "gene", cursor)
gene_novelty_types = [ x[-2] for x in annotation['gene_novelty']]
t_novelty_types = [ x[-2] for x in annotation['transcript_novelty']]
assert annotation['gene_novelty'][0][-1] == "TRUE"
assert "antisense_gene" in gene_novelty_types
assert "antisense_transcript" in t_novelty_types
conn.close()
