diff --git a/src/intronIC/extraction/sequences.py b/src/intronIC/extraction/sequences.py index c5cd403..bf57a9b 100644 --- a/src/intronIC/extraction/sequences.py +++ b/src/intronIC/extraction/sequences.py @@ -365,41 +365,12 @@ class SequenceExtractor: # Extract branch point region # BP coords are relative to 3' splice site (acceptor) # Adjust to be 0-based and relative to intron end - bp_start_rel = bp_coords[0] # Negative value (e.g., -55) - bp_end_rel = bp_coords[1] # Negative value (e.g., -5) - - if coord.strand == "+": - # For + strand: 3' splice site is at coord.stop - # bp_start_rel is more negative (e.g., -55), so further upstream - # bp_end_rel is less negative (e.g., -5), so closer to 3'ss - bp_start_abs = coord.stop + bp_start_rel - 1 # -1 for 0-based - bp_end_abs = coord.stop + bp_end_rel - 1 - else: - # For - strand: 3' splice site is at coord.start - # On - strand, the coordinates are reversed: more negative means further - # in genomic coords (higher position), so we swap start/end - # This matches v1.5.1: bpc = (bp_stop, bp_start) for reverse strand - bp_end_abs = coord.start - bp_start_rel - 1 # Note: assigned to end - bp_start_abs = coord.start - bp_end_rel - 1 # Note: assigned to start - - # Extract BP region (different handling for cached vs indexed mode) - if region_seq is not None: - # Cached mode: slice from region_seq - bp_start_abs = max(0, bp_start_abs) - bp_end_abs = min(len(region_seq), bp_end_abs) - bp_region_seq = region_seq[bp_start_abs:bp_end_abs] - else: - # Indexed mode: fetch from genome - # Convert back to 1-based for fetch - bp_fetch_start = max(1, bp_start_abs + 1) - bp_fetch_stop = bp_end_abs + 1 - - bp_region_seq = self.genome_reader.fetch( - coord.chromosome, bp_fetch_start, bp_fetch_stop - ).upper() - - if coord.strand == "-": - bp_region_seq = reverse_complement(bp_region_seq) + # intron_seq is already in transcript orientation on both strands. + # Use half-open offsets from its end, without a second coordinate + # conversion or an inclusive indexed fetch. + bp_start = max(0, len(intron_seq) + bp_coords[0]) + bp_stop = max(0, min(len(intron_seq), len(intron_seq) + bp_coords[1])) + bp_region_seq = intron_seq[bp_start:bp_stop] # Identify terminal dinucleotides five_prime_dnt = intron_seq[:2] if len(intron_seq) >= 2 else intron_seq