digraph G {
  compound=true;
  CleanData -> Chemical;
  Chemical [label = "get CHEMICAL medicine type"];
  CleanData -> PEmap;
  PEmap [label = "paired reads mapping"];
  raw->mapping->bam_filter->var_calling->var_filter->result[arrowhead=none];
  PEmap -> DEunpair;
  Rmdup1 [label = "rmdup for fusion"];
  Realign [label = "recal && realign"];
  Target [label = "get targeted bam"];
  DEunpair [label = "delete unmapped or \n unpaired reads"];
  FilterOther[fillcolor=yellow, style="rounded,filled", shape=box];
  Filter[fillcolor=yellow, style="rounded,filled", shape=box];
  STAT[fillcolor=yellow, style="rounded,filled", shape=box];
  Filter1[fillcolor=yellow, style="rounded,filled", shape=box];
  Filter2[fillcolor=yellow, style="rounded,filled", shape=box];
  Filter_fusion1[fillcolor=yellow, style="rounded,filled", shape=box];
  ANNO_fusion[shape=box];
  CHEMICAL_medicine_Type[fillcolor=yellow, style="rounded,filled", shape=diamond, shape=box];
  RmXA [label = "rm Multiple alignment"];
  FilterOther [label ="filter no md or cigar\nfilter much mismatch:\ninsersion and deletion >= 3\nall type mismatch > 5\nsoft clip and border 5bp uncount"];
  MergeDP [label = "merge business and control depth"];
  Filter [label = "filter cnv by gene:\nmore than 70% base of one gene > 0.7\nbase of one gene detected base > 500"];
  Chrcalling [label = "calling each chr snv by mutect2"];
  Merge [label = "merge vcf"];
  Add [label ="add cosmic and rs data"];
  STAT [label = "stat mismatch had high quality:\nmapping quality >=30\ninsersion and deletion < 2\nall type mismatch <= 3\nborder > 5bp\nbase quality >=20 or 80% nearby 10bp base quality >= 20"];
  Filter1 [label = "filter depth low:\ndepth all one base >= 30 \ndepth of alt >=5\nalt ratio >= 0.002"];
  Filter2 [label = "get mismatch had high quality:\n1.save the snv supported by >= 5 differ type reads\n and >= 1 positive read >= 1 reverse read\n2.indel have no high quality reads support tagged as non-filter\n3.snp have no high quality reads support tagged as filtered"];
  FUSION [label = "CREST detect SV raw result"];
  Filter_fusion1 [label = "filter:\none or more percent_identity < 0.9\none or more depth < 4\nboth side out of target +-50bp"];
  POS [label = "get the postion of both split side"];
  ANNO_fusion [label = "annovar for the both side split pos\n to get gene information"];
  merge_result [label = "merge the annovar result to sv result"];
  subgraph cluster0 {
    label = "mapping bwa";
    DEunpair -> Sorted;
    Sorted -> Target;
    Target -> Realign;
    Sorted -> Rmdup1;
    Realign -> Rmdup;
    Rmdup1 -> "sample.rmdup.bam"[splines="line"];
    Realign -> "sample.target.realigned.rmxa.bam";
    {rank=same;  Rmdup, Realign};
  }
  subgraph cluster1 {
    label = "bam filter";
    RmXA -> FilterOther;
    FilterOther -> "sample.target.realigned.removemultiple.bam";
  }
  subgraph cluster3{
    label = "CNV calling";
    "sample.target.bam.depth" -> MergeDP;
    MergeDP -> CNVcalling;
    CNVcalling -> Filter;
  }
  subgraph cluster4{
    label = "SNV calling";
    Chrcalling -> Merge;
    Merge -> Add;
    Add -> ANNO;
  }
  subgraph cluster5{
    label = "somatic SNV filter";
    STAT -> Filter1;
    Filter1 -> Filter2;
    Filter2 -> "sample.somtic.new.xlsx";
  }
  subgraph cluster6{
    label = "FUSION detect by CREST";
    FUSION -> Filter_fusion1;
    Filter_fusion1 -> POS;
    POS -> ANNO_fusion;
    ANNO_fusion -> merge_result;
  }
  splines=ortho;
  CHEMICAL_medicine_Type [label = "CHEMICAL medicine Type:\nhom_alt alt_ratio >= 0.8\nhet alt_ratio >= 0.3\nother hom_ref\nthe ssr is not detect,\n one of it is typed by indel"];
  "sample.rmdup.bam" -> FUSION;
  "sample.target.realigned.rmxa.bam" -> RmXA;
  "sample.target.realigned.removemultiple.bam" -> Chrcalling;
  Target -> "sample.target.bam.depth";
  ANNO -> STAT;
  Chemical -> CHEMICAL_medicine_Type;
  Merge -> CHEMICAL_medicine_Type;
  CHEMICAL_medicine_Type -> "sample.CHEMICAL.medicine.xlsx";
  merge_result -> "fusion.result.xlsx";
  Filter -> "cnv.xlsx";
  "cnv.xlsx"[fillcolor=green, style="rounded,filled", shape=box];
  "sample.somtic.new.xlsx"[fillcolor=green, style="rounded,filled", shape=box];
  "sample.CHEMICAL.medicine.xlsx"[fillcolor=green, style="rounded,filled", shape=box];
  "fusion.result.xlsx"[fillcolor=green, style="rounded,filled", shape=box];
  "sample.target.realigned.rmxa.bam"[fillcolor=pink, style="filled",shape=box];
  "sample.target.realigned.removemultiple.bam"[fillcolor=pink, style="filled",shape=box];
  "sample.target.bam.depth"[fillcolor=pink, style="filled",shape=box];
  "sample.rmdup.bam"[fillcolor=pink, style="filled",shape=box];
  { rank=same; raw; CleanData;}
  { rank=same; bam_filter; "sample.target.realigned.rmxa.bam", "sample.target.bam.depth";}
  { rank=same; var_calling; "sample.rmdup.bam","sample.target.realigned.removemultiple.bam";}
  { rank=same; var_filter;  FUSION;}
  { rank=same; result; "sample.somtic.new.xlsx","sample.CHEMICAL.medicine.xlsx","fusion.result.xlsx","cnv.xlsx";}
}