Commit 3fedaecf authored by Ivan Vanat's avatar Ivan Vanat
Browse files

merging trfs with overlap

parent 18588449

nested/core/tools/tandem_repeat_finder.py

+17 −0
Original line number Diff line number Diff line
@@ -67,6 +67,23 @@ class TandemRepeatFinder(ToolInterface): 
        result = []

        candidates.sort(key=lambda repeat: repeat.location[0])



        for candidate in candidates:

            if not result:

                result.append(candidate)

            else:

                last = result[-1]

                if intervals.intersect(last.location, candidate.location):

                    right = max(last.location[1], candidate.location[1])

                    last.location[1] = right

                else:

                    result.append(candidate)



        return result



    def _filter_candidates_assign_overlap_strategy(self, candidates: List[TandemRepeat]) -> List[TandemRepeat]:

        result = []

        candidates.sort(key=lambda repeat: repeat.location[0])



        for candidate in candidates:

            if not result:

                result.append(candidate)

nested/output/output_generators/trf_output_generator.py

+6 −6
Original line number Diff line number Diff line
@@ -13,8 +13,7 @@ class TrfOutputGenerator(OutputGeneratorInterface): 
    def generate_output(self, output_object: TrfOutputObject) -> NoReturn:

        self.create_data_directory(output_object.sequence_id)



        with open(f"{self.directory}/{output_object.sequence_id}/{output_object.sequence_id}_trf.gff", "w+")\

                as gff:

        with open(f"{self.directory}/{output_object.sequence_id}/{output_object.sequence_id}_trf.gff", "w+") as gff:

            i = 0

            gff.write("##gff-version 3\n")

            for repeat in output_object.elements:
@@ -23,9 +22,10 @@ class TrfOutputGenerator(OutputGeneratorInterface): 
                          + str(repeat.location[0]) + "\t"

                          + str(repeat.location[1]) + "\t"

                          + ".\t.\t.\t"

                          + "ID=tandem_repeat_" + str(i)

                          + ";monomer_count=" + str(repeat.copies)

                          + ";monomer=" + repeat.monomer

                          + ";color=" + config["igv_colors"]["trf"][repeat.type]

                          # + "ID=tandem_repeat_" + str(i)

                          # + ";monomer_count=" + str(repeat.copies)

                          # + ";monomer=" + repeat.monomer

                          # + ";color=" + config["igv_colors"]["trf"][repeat.type]

                          + "."

                          + "\n")

                i += 1