def test_FSM_end_diff(self): """ Example where the transcript is an FSM but has a difference on the ends large enough to be novel. """ 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) 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, gene_ends = init_refs.make_gene_start_and_end_dict( cursor, build) chrom = "chr2" strand = "+" positions = [1, 100, 500, 600, 900, 1500] annotation = talon.identify_transcript(chrom, positions, strand, cursor, location_dict, edge_dict, transcript_dict, vertex_2_gene, gene_starts, gene_ends, run_info, "temp_gene") correct_gene_ID = fetch_correct_ID("TG2", "gene", cursor) novelty_types = [x[-2] for x in annotation['transcript_novelty']] assert annotation['gene_ID'] == correct_gene_ID assert annotation['end_delta'] == None 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, gene_ends = init_refs.make_gene_start_and_end_dict( cursor, build) 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()
def test_genomic_unspliced(self): """ Monoexonic fragment that overlaps gene 1 """ 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, gene_ends = init_refs.make_gene_start_and_end_dict( cursor, build) chrom = "chr1" strand = "+" positions = (1, 990) 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("TG1", "gene", cursor) novelty_types = [x[-2] for x in annotation['transcript_novelty']] assert annotation['gene_ID'] == correct_gene_ID assert "genomic_transcript" in novelty_types assert annotation['end_delta'] == -10 conn.close()
def test_vertex2gene_update(self): """ Update vertex to gene relationships """ conn, cursor = get_db_cursor() build = "toy_build" vertex_2_gene = init_refs.make_vertex_2_gene_dict(cursor) talon.update_vertex_2_gene(2, (1, 2), "-", vertex_2_gene) talon.update_vertex_2_gene(1, (1, 2, 3, 4, 5, 6), "+", vertex_2_gene) # Write to file os.system("mkdir -p scratch/db_updates/") with open("scratch/db_updates/v2g.tsv", 'w') as f: for vertex_ID, gene_set in vertex_2_gene.items(): for gene in gene_set: entry = "\t".join([str(x) for x in (vertex_ID, gene[0])]) f.write(entry + "\n") batch_size = 100 talon.batch_add_vertex2gene(cursor, "scratch/db_updates/v2g.tsv", batch_size) # Use queries to check if the insert worked as expected query = "SELECT * FROM vertex WHERE vertex_ID = '1'" cursor.execute(query) gene_IDs = [x['gene_ID'] for x in cursor.fetchall()] assert gene_IDs == [1, 2] query = "SELECT * FROM vertex WHERE gene_ID = '1'" cursor.execute(query) vertex_IDs = [x['vertex_ID'] for x in cursor.fetchall()] assert vertex_IDs == [1, 2, 3, 4, 5, 6]
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_NNC(self): """ Example where the transcript skips an exon and has a novel splice donor """ 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) 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, gene_ends = init_refs.make_gene_start_and_end_dict( cursor, build) chrom = "chr1" strand = "+" positions = [1, 50, 900, 1000] annotation = talon.identify_transcript(chrom, positions, strand, cursor, location_dict, edge_dict, transcript_dict, vertex_2_gene, gene_starts, gene_ends, run_info, "temp_gene") correct_gene_ID = fetch_correct_ID("TG1", "gene", cursor) novelty_types = [x[-2] for x in annotation['transcript_novelty']] assert annotation['gene_ID'] == correct_gene_ID assert "NNC_transcript" in novelty_types assert annotation['start_delta'] == annotation['end_delta'] == 0 conn.close()
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_FSM_perfect(self): """ Example where the transcript is a perfect full splice 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) 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, gene_ends = init_refs.make_gene_start_and_end_dict( cursor, build) chrom = "chr1" strand = "+" positions = [1, 100, 500, 600, 900, 1000] annotation = talon.identify_transcript(chrom, positions, strand, cursor, location_dict, edge_dict, transcript_dict, vertex_2_gene, gene_starts, gene_ends, run_info, "temp_gene") correct_gene_ID = fetch_correct_ID("TG1", "gene", cursor) correct_transcript_ID = fetch_correct_ID("TG1-001", "transcript", cursor) assert annotation['gene_ID'] == correct_gene_ID assert annotation['transcript_ID'] == correct_transcript_ID assert annotation['transcript_novelty'] == [] conn.close()
def test_NIC_instead_of_ISM(self): """ Test case where the transcript looks like an ISM, but is NIC on account of having known starts and ends """ 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) 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, gene_ends = init_refs.make_gene_start_and_end_dict( cursor, build) chrom = "chr3" strand = "+" positions = (800, 1000, 1200, 1400, 1600, 1800, 2000, 2200) annotation = talon.identify_transcript(chrom, positions, strand, cursor, location_dict, edge_dict, transcript_dict, vertex_2_gene, gene_starts, gene_ends, run_info, "tmp_gene") correct_gene_ID = fetch_correct_ID("TG5", "gene", cursor) novelty_types = [x[-2] for x in annotation['transcript_novelty']] assert annotation['gene_ID'] == correct_gene_ID assert "NIC_transcript" in novelty_types conn.close()
def test_all_v2g(self): """ Get all vertices in the database """ conn, cursor = get_db_cursor() build = "toy_build" v2g_dict = init_refs.make_vertex_2_gene_dict(cursor, build) conn.close() assert len(v2g_dict.keys()) == 34
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_interval_v2g(self): """ Get all vertices in the database within the specified interval """ conn, cursor = get_db_cursor() build = "toy_build" v2g_dict = init_refs.make_vertex_2_gene_dict(cursor, build, chrom="chr1", start=1, end=1000) conn.close() assert len(v2g_dict.keys()) == 6
def test_no_match(self): """ Example where no match exists """ conn, cursor = get_db_cursor() db = "scratch/toy.db" build = "toy_build" init_refs.make_temp_novel_gene_table(cursor, "toy_build") run_info = talon.init_run_info(db, build) vertex2gene = init_refs.make_vertex_2_gene_dict(cursor) vertex_IDs = (1000, 2000, 3000, 4000) strand = "+" gene_ID = talon.find_gene_match_on_vertex_basis(vertex_IDs, strand, vertex2gene) assert gene_ID == None 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_perfect_match(self): """ Example where the vertices perfectly match a gene. """ conn, cursor = get_db_cursor() db = "scratch/toy.db" build = "toy_build" init_refs.make_temp_novel_gene_table(cursor, "toy_build") run_info = talon.init_run_info(db, build) vertex2gene = init_refs.make_vertex_2_gene_dict(cursor) vertex_IDs = (1, 2, 3, 4, 5, 6) strand = "+" gene_ID = talon.find_gene_match_on_vertex_basis(vertex_IDs, strand, vertex2gene) correct_gene_ID = fetch_correct_ID("TG1", "gene", cursor) assert gene_ID == correct_gene_ID conn.close()
def test_NIC_with_all_known_edges(self): """ Test case derived from a real mouse Map2k4 read. All of edges are known (except 3'), yet the read is NIC not FSM/ISM """ database = "scratch/Map2k4.db" talon.get_counters(database) conn = sqlite3.connect(database) conn.row_factory = sqlite3.Row cursor = conn.cursor() build = "mm10" init_refs.make_temp_novel_gene_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, gene_ends = init_refs.make_gene_start_and_end_dict( cursor, build) chrom = "chr11" strand = "-" positions = [ 65788254, 65788136, 65775765, 65775733, 65756371, 65756269, 65735366, 65735192, 65719603, 65719484, 65712297, 65712178, 65709983, 65709932, 65707111, 65706984, 65696365, 65696288, 65693570, 65693422, 65691773, 65691728, 65690804, 65689322 ] annotation = talon.identify_transcript(chrom, positions, strand, cursor, location_dict, edge_dict, transcript_dict, vertex_2_gene, gene_starts, gene_ends, run_info, "temp_gene") assert annotation['gene_ID'] == 1 assert annotation['transcript_ID'] == 8 novelty_types = [x[-2] for x in annotation['transcript_novelty']] assert "NIC_transcript" in novelty_types 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_spliced_antisense(self): """ Example where the transcript matches known vertices but 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) 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, gene_ends = init_refs.make_gene_start_and_end_dict( cursor, build) chrom = "chr2" strand = "-" positions = [1000, 900, 600, 500, 100, 1] annotation = talon.identify_transcript(chrom, positions, strand, cursor, location_dict, edge_dict, transcript_dict, vertex_2_gene, gene_starts, gene_ends, run_info, "temp_gene") anti_gene_ID = fetch_correct_ID("TG2", "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 assert annotation['start_delta'] == annotation['end_delta'] == 0 conn.close()