diff --git a/bin/generate_exint_plotter_input.py b/bin/generate_exint_plotter_input.py --- a/bin/generate_exint_plotter_input.py +++ b/bin/generate_exint_plotter_input.py @@ -24,6 +24,11 @@ my_ref_prots_file = args.ref_prots my_species = args.species my_output_file = args.output_file + +def coord_key(exon_id): + #numeric sort key for "chr:start-stop" (a plain string sort puts 100500 before 99000) + m = re.match(r"^(.*):(\d+)-(\d+)$", exon_id) + return (m.group(1), int(m.group(2)), int(m.group(3))) if m else (exon_id, 0, 0) ################ OVERLAPPING EXONS INFO ##################### my_overlap_df = pd.read_table(my_overlap_file, sep="\t", header=None, names=["ExOverlapID", "GeneID", "ExCoords","Strand"]) @@ -143,7 +148,7 @@ UpDown_phases_df = pd.DataFrame(columns=["GeneID", "ExonID", "Strand", "UpPhase", "DownPhase"]) my_grouped_df = my_phase_strand_df.groupby("GeneID") #for each gene for name, group in my_grouped_df: - gene_exons_dict[name] = sorted(list(group["ExonID"])) #dict value = sorted list of exons for each gene (key) + gene_exons_dict[name] = sorted(list(group["ExonID"]), key=coord_key) #dict value = sorted list of exons for each gene (key) for my_exon in list(my_phase_strand_df.loc[:,"ExonID"]): #for each exon my_gene = exon_gene_dict[my_exon] #select the gene @@ -227,9 +232,9 @@ #define the functions def reorder_exons(group): if list(group["Strand"])[0] == "+": #check first element of strand list. They should be all the same anyways. - my_group = group.sort_values(by=["ExonID"]) #order in ascending order + my_group = group.iloc[sorted(range(group.shape[0]), key=lambda i: coord_key(list(group["ExonID"])[i]))] #order in ascending order elif list(group["Strand"])[0] == "-": - my_group = group.sort_values(by=["ExonID"], ascending=False) + my_group = group.iloc[sorted(range(group.shape[0]), key=lambda i: coord_key(list(group["ExonID"])[i]), reverse=True)] return(my_group) def add_fake_coords(group):