build_splicegraph.py 文件源码

def compute_edge_overlaps(splice_graph):
    """Get the overlap between connected exons:
    - Positive overlap means that they overlap that number of bases,
    - Zero that they occur next to each other
    - Negative that there is a gap in the transcriptome of that number of bases (one or multiple exons of length < kmer)

    Note: the splice graph must have already the nodes written with coordinates, and the edges alredy entered too.
    """
    #Init
    edge_overlaps = {edge: None for edge in splice_graph.edges()}
    exon2coord = nx.get_node_attributes(
        G=splice_graph,
        name='coordinates'
    )

    for (node1, node2) in sorted(edge_overlaps.keys()):

        # Get the list of transcripts that they belong
        node1_transcripts = set(coordinate[0] for coordinate  in exon2coord[node1])
        node2_transcripts = set(coordinate[0] for coordinate  in exon2coord[node2])
        intersection = node1_transcripts & node2_transcripts
        a_common_transcript = intersection.pop()

        # Get the end the first
        node1_coords = exon2coord[node1]
        node1_coords_in_transcript = [x for x in node1_coords if x[0] == a_common_transcript][0]
        node1_end = node1_coords_in_transcript[2]

        # Get the start of the next
        node2_coords = exon2coord[node2]
        node2_coords_in_transcript = [x for x in node2_coords if x[0] == a_common_transcript][0]
        node2_start = node2_coords_in_transcript[1]

        # Overlap in bases, 0 means one next to the other, negative numbers a gap
        overlap = node1_end - node2_start
        edge_overlaps[(node1, node2)] = overlap

    return edge_overlaps