[partition] Do not run clustering when k = 1

cmd/allhic.go

@@ -162,12 +162,16 @@ can be generated with the "extract" sub-command.
 			Name:  "optimize",
 			Usage: "Order-and-orient tigs in a group",
 			UsageText: `
-	allhic optimize counts_RE.txt clmfile [options]
+	allhic optimize counts_RE.txt clmfile clusters.txt group_number [options]
 
 Optimize function:
 Given a set of Hi-C contacts between contigs, as specified in the
 clmfile, reconstruct the highest scoring ordering and orientations
-for these contigs.
+for these contigs. Optimize run on a specific partition in "clusters.txt"
+as generated by "partition" sub-command, with the group_number matching the
+order appearing in "clusters.txt". Typically, if there are k clusters, we
+can start k separate "optimize" commands for parallelism (for example,
+on a cluster).
 `,
 			Flags: []cli.Flag{
 				cli.BoolFlag{
@@ -200,20 +204,23 @@ for these contigs.
 				},
 			},
 			Action: func(c *cli.Context) error {
-				if len(c.Args()) < 2 {
+				if len(c.Args()) < 4 {
 					cli.ShowSubcommandHelp(c)
 					return cli.NewExitError("Must specify clmfile", 1)
 				}
 
 				refile := c.Args().Get(0)
 				clmfile := c.Args().Get(1)
+				clustersfile := c.Args().Get(2)
+				group, _ := strconv.Atoi(c.Args().Get(3))
 				runGA := !c.Bool("skipGA")
 				startOver := c.Bool("startOver")
 				seed := c.Int64("seed")
 				npop := c.Int("npop")
 				ngen := c.Int("ngen")
 				mutpb := c.Float64("mutpb")
 				p := allhic.Optimizer{REfile: refile, Clmfile: clmfile,
+					Clustersfile: clustersfile, Group: group,
 					RunGA: runGA, StartOver: startOver,
 					Seed: seed, NPop: npop, NGen: ngen, MutProb: mutpb}
 				p.Run()

optimize.go

@@ -19,15 +19,17 @@ import (
 
 // Optimizer runs the order-and-orientation procedure, given a clmfile
 type Optimizer struct {
-	Clmfile   string
-	REfile    string
-	RunGA     bool
-	StartOver bool
-	Seed      int64
-	NPop      int
-	NGen      int
-	MutProb   float64
-	CrossProb float64
+	Clmfile      string
+	REfile       string
+	Clustersfile string
+	Group        int
+	RunGA        bool
+	StartOver    bool
+	Seed         int64
+	NPop         int
+	NGen         int
+	MutProb      float64
+	CrossProb    float64
 }
 
 // Run kicks off the Optimizer

partition.go

@@ -30,15 +30,34 @@ type Partitioner struct {
 // Run is the main function body of partition
 func (r *Partitioner) Run() {
 	r.readRE()
-	r.makeMatrix()
 	r.skipContigsWithFewREs()
-	r.skipRepeats()
-	r.Cluster()
+	if r.K == 1 {
+		r.makeTrivialClusters()
+	} else {
+		r.makeMatrix()
+		r.skipRepeats()
+		r.Cluster()
+	}
 	r.printClusters()
-
 	log.Notice("Success")
 }
 
+// makeTrivialClusters make a single cluster containing all contigs
+// except the really short ones
+func (r *Partitioner) makeTrivialClusters() {
+	contigs := []int{}
+	for i, contig := range r.contigs {
+		if contig.skip {
+			continue
+		}
+		contigs = append(contigs, i)
+	}
+	clusters := Clusters{
+		0: contigs,
+	}
+	r.clusters = clusters
+}
+
 // skipContigsWithFewREs skip contigs with fewere than MinREs
 // This reads in the `counts_RE.txt` file generated by extract()
 func (r *Partitioner) skipContigsWithFewREs() {
@@ -160,8 +179,8 @@ func (r *Partitioner) readRE() {
 	for i, contig := range r.contigs {
 		r.contigToIdx[contig.name] = i
 	}
-	log.Noticef("Loading contig RE lengths for normalization from `%s`",
-		r.Contigsfile)
+	log.Noticef("Loaded %d contig RE lengths for normalization from `%s`",
+		len(r.contigs), r.Contigsfile)
 }
 
 // parseDist imports the edges of the contig into a slice of DistLine