From c36c4f48f9c028e24ecc4d7f3aba1f90b6aa0688 Mon Sep 17 00:00:00 2001
From: Haibao Tang <tanghaibao@gmail.com>
Date: Mon, 3 Sep 2018 16:25:08 -0700
Subject: [PATCH] [extract] Add label column to pairs.txt

---
 extract.go   | 9 +++++----
 partition.go | 2 ++
 prune.go     | 5 +++++
 3 files changed, 12 insertions(+), 4 deletions(-)

diff --git a/extract.go b/extract.go
index 0b7d2e0..6fce4f4 100644
--- a/extract.go
+++ b/extract.go
@@ -64,6 +64,7 @@ type ContigPair struct {
 	L1, L2         int
 	nObservedLinks int
 	nExpectedLinks float64
+	label          string // allelic/cross-allelic/ok
 }
 
 // String outputs the string representation of ContigInfo
@@ -73,9 +74,9 @@ func (r ContigInfo) String() string {
 
 // String outputs the string representation of ContigInfo
 func (r ContigPair) String() string {
-	return fmt.Sprintf("%d\t%d\t%s\t%s\t%d\t%d\t%d\t%.1f",
+	return fmt.Sprintf("%d\t%d\t%s\t%s\t%d\t%d\t%d\t%.1f\t%s",
 		r.ai, r.bi, r.at, r.bt, r.RE1, r.RE2,
-		r.nObservedLinks, r.nExpectedLinks)
+		r.nObservedLinks, r.nExpectedLinks, r.label)
 }
 
 // uintLog2 calculates the integer log2 of a number
@@ -241,7 +242,7 @@ func (r *Extracter) calcInterContigs() {
 			L1, L2 := ca.length, cb.length
 			cp = &ContigPair{ai: ai, bi: bi, at: at, bt: bt,
 				RE1: ca.recounts, RE2: cb.recounts,
-				L1: L1, L2: L2}
+				L1: L1, L2: L2, label: "ok"}
 			cp.nExpectedLinks = sumf(r.findExpectedInterContigLinks(0, L1, L2))
 			cp.nObservedLinks = len(line.links)
 			contigPairs[pair] = cp
@@ -253,7 +254,7 @@ func (r *Extracter) calcInterContigs() {
 	f, _ := os.Create(outfile)
 	w := bufio.NewWriter(f)
 	defer f.Close()
-	fmt.Fprintf(w, "#X\tY\tContig1\tContig2\tRE1\tRE2\tObservedLinks\tExpectedLinksIfAdjacent\n")
+	fmt.Fprintf(w, "#X\tY\tContig1\tContig2\tRE1\tRE2\tObservedLinks\tExpectedLinksIfAdjacent\tLabel\n")
 
 	allPairs := []*ContigPair{}
 	for _, c := range contigPairs {
diff --git a/partition.go b/partition.go
index b1e4b31..8561b38 100644
--- a/partition.go
+++ b/partition.go
@@ -237,12 +237,14 @@ func parseDist(pairsFile string) []ContigPair {
 		RE2, _ := strconv.Atoi(rec[5])
 		nObservedLinks, _ := strconv.Atoi(rec[6])
 		nExpectedLinks, _ := strconv.ParseFloat(rec[7], 64)
+		label := rec[8]
 
 		cp := ContigPair{
 			ai: ai, bi: bi,
 			at: at, bt: bt,
 			RE1: RE1, RE2: RE2,
 			nObservedLinks: nObservedLinks, nExpectedLinks: nExpectedLinks,
+			label: label,
 		}
 
 		edges = append(edges, cp)
diff --git a/prune.go b/prune.go
index 161d228..3694cae 100644
--- a/prune.go
+++ b/prune.go
@@ -26,6 +26,11 @@ type Pruner struct {
 type AlleleGroup []string
 
 // Run calls the pruning steps
+// The pruning algorithm is a heuristic method that removes the following pairs:
+//
+// 1. Alleleic, these are directly the pairs of allelic contigs given in the allele table
+// 2. Cross-allelic, these are any contigs that connect to the allelic contigs so we only keep the
+//    the best pair
 func (r *Pruner) Run() {
 	edges := parseDist(r.PairsFile)
 	alleleGroups := parseAllelesTable(r.AllelesFile)