func main() {
flag.Parse()
w := gff.NewWriter(os.Stdout, 60, false)
sc := featio.NewScanner(gff.NewReader(os.Stdin))
for sc.Next() {
f := sc.Feat().(*gff.Feature)
r := f.FeatAttributes.Get("Repeat")
fields := strings.Fields(r)
if len(fields) < 4 {
log.Fatal("invalid repeat attribute")
}
end, err := strconv.Atoi(fields[3])
if err != nil {
log.Fatalf("failed to parse end coordinate: %v", err)
}
remainder, err := strconv.Atoi(fields[4])
if err != nil {
log.Fatalf("failed to parse remains coordinate: %v", err)
}
length := end + remainder
if length < *thresh {
continue
}
w.Write(f)
}
if err := sc.Error(); err != nil {
log.Fatalf("error during gff read: %v", err)
}
}
func readAnnotations(file string) (map[string]*interval.IntTree, error) {
f, err := os.Open(file)
if err != nil {
return nil, err
}
trees := make(map[string]*interval.IntTree)
sc := featio.NewScanner(gff.NewReader(f))
for id := uintptr(1); sc.Next(); id++ {
f := sc.Feat().(*gff.Feature)
t, ok := trees[f.SeqName]
if !ok {
t = &interval.IntTree{}
trees[f.SeqName] = t
}
t.Insert(gffInterval{f, id}, true)
}
err = sc.Error()
if err != nil {
log.Fatalf("error during GFF read: %v", err)
}
for _, t := range trees {
t.AdjustRanges()
}
return trees, nil
}
func main() {
flag.Parse()
if *in == "" {
flag.Usage()
os.Exit(1)
}
f, err := os.Open(*in)
if err != nil {
log.Fatalf("failed to open %q: %v", *in, err)
}
events := make(map[string][]*gff.Feature)
fsc := featio.NewScanner(gff.NewReader(f))
for fsc.Next() {
f := fsc.Feat().(*gff.Feature)
fields := strings.Fields(f.FeatAttributes.Get("Read"))
if len(fields) != 3 {
log.Fatalf("bad record: %+v", f)
}
events[fields[0]] = append(events[fields[0]], f)
}
if err := fsc.Error(); err != nil {
log.Fatalf("error during gff read: %v", err)
}
f.Close()
for _, ref := range flag.Args() {
f, err = os.Open(ref)
if err != nil {
log.Fatalf("failed to open reference %q: %v", ref, err)
}
ssc := seqio.NewScanner(fasta.NewReader(f, linear.NewSeq("", nil, alphabet.DNA)))
for ssc.Next() {
seq := ssc.Seq().(*linear.Seq)
for _, f := range events[seq.Name()] {
fields := strings.Fields(f.FeatAttributes.Get("Read"))
if len(fields) != 3 {
log.Fatalf("bad record: %+v", f)
}
start, err := strconv.Atoi(fields[1])
if err != nil {
log.Fatalf("failed to get start coordinate: %v", err)
}
end, err := strconv.Atoi(fields[2])
if err != nil {
log.Fatalf("failed to get end coordinate: %v", err)
}
tmp := *seq
tmp.ID += fmt.Sprintf("//%d_%d", start, end)
tmp.Seq = tmp.Seq[start:end]
fmt.Printf("%60a\n", &tmp)
}
}
if err := ssc.Error(); err != nil {
log.Fatalf("error during fasta read: %v", err)
}
f.Close()
}
}
func filterFeats(annot string, classes, names []string, thresh float64) ([]interval.IntTree, error) {
ntab := make(map[string]int, len(names))
for i, n := range names {
ntab[n] = i
}
ts := make([]interval.IntTree, len(names))
cm := make(map[string]struct{})
for _, c := range classes {
cm[c] = struct{}{}
}
f, err := os.Open(annot)
if err != nil {
return nil, err
}
fs := featio.NewScanner(gff.NewReader(f))
for fs.Next() {
f := fs.Feat().(*gff.Feature)
if f.FeatScore == nil || math.Exp(*f.FeatScore) < thresh {
continue
}
var class string
if att := f.FeatAttributes.Get(f.Feature); att != "" {
// This gets the repeat attributes only.
class = f.Feature + "/" + strings.Fields(att)[1]
} else {
class = f.Feature
}
if _, ok := cm[class]; !ok {
last := strings.LastIndex(class, "/")
if last == strings.Index(class, "/") {
continue
}
if _, ok := cm[class[:last]]; !ok {
continue
}
}
if chr, ok := ntab[f.SeqName]; ok {
ts[chr].Insert(intGff{f}, true)
}
}
if err := fs.Error(); err != nil {
return nil, err
}
for i := range ts {
ts[i].AdjustRanges()
}
return ts, nil
}
func annotFeats(annot string, classes, names []string) ([]interval.IntTree, error) {
ntab := make(map[string]int, len(names))
for i, n := range names {
ntab[n] = i
}
ts := make([]interval.IntTree, len(names))
cm := make(map[string]struct{})
for _, c := range classes {
cm[c] = struct{}{}
}
f, err := os.Open(annot)
if err != nil {
return nil, err
}
fs := featio.NewScanner(gff.NewReader(f))
for id := uintptr(0); fs.Next(); id++ {
f := fs.Feat().(*gff.Feature)
var class string
att := f.FeatAttributes.Get("repeat")
if att == "" {
// Ignore non-repeat features.
continue
}
repeatFields := strings.Fields(att)
class = f.Feature + "/" + repeatFields[1]
if _, ok := cm[class]; !ok {
last := strings.LastIndex(class, "/")
if last == strings.Index(class, "/") {
continue
}
if _, ok := cm[class[:last]]; !ok {
continue
}
}
if chr, ok := ntab[f.SeqName]; ok {
ts[chr].Insert(intGff{f, id}, true)
}
}
if err := fs.Error(); err != nil {
return nil, err
}
for i := range ts {
ts[i].AdjustRanges()
}
return ts, nil
}
func main() {
flag.Parse()
if *exclude == "" {
flag.Usage()
os.Exit(1)
}
nameSet := make(map[string]struct{})
f, err := os.Open(*exclude)
if err != nil {
log.Fatalf("failed to open exclude file %q: %v", *exclude, err)
}
ls := bufio.NewScanner(f)
for ls.Scan() {
nameSet[ls.Text()] = struct{}{}
}
err = ls.Err()
if err != nil {
log.Fatalf("failed to read exclude file: %v", err)
}
w := gff.NewWriter(os.Stdout, 60, true)
var excl *gff.Writer
if *retain {
excl = gff.NewWriter(os.Stderr, 60, true)
}
sc := featio.NewScanner(gff.NewReader(os.Stdin))
for sc.Next() {
f := sc.Feat().(*gff.Feature)
n := f.FeatAttributes.Get("Read")
if _, ok := nameSet[n]; ok {
if excl != nil {
_, err := excl.Write(f)
if err != nil {
log.Fatalf("failed to write feature: %v", err)
}
}
continue
}
_, err := w.Write(f)
if err != nil {
log.Fatalf("failed to write feature: %v", err)
}
}
if err := sc.Error(); err != nil {
log.Fatalf("error during gff read: %v", err)
}
}
func main() {
flag.Parse()
var grps []map[string]int
sc := featio.NewScanner(gff.NewReader(os.Stdin))
for sc.Next() {
f := sc.Feat().(*gff.Feature)
r := f.FeatAttributes.Get("Repeat")
g := f.FeatAttributes.Get("Group")
typ := strings.Fields(r)[0]
if !*doGrouping {
fmt.Printf("%s\t%s\n", g, typ)
}
gid, err := strconv.Atoi(g)
if err != nil {
log.Fatalf("failed to parse group id: %v", err)
}
grps = add(grps, gid, typ)
}
if err := sc.Error(); err != nil {
log.Fatalf("error during gff read: %v", err)
}
if !*doGrouping {
return
}
for gid, g := range grps {
if g == nil {
continue
}
fmt.Printf("%d\t", gid)
m := sortedMap(g)
for i, t := range m {
if i != 0 {
fmt.Print(" ")
}
fmt.Printf("%s:%d", t.typ, t.n)
}
name := nameHeuristic(m)
fmt.Printf("\t%s\t%s\n", name, trunc(name, 5))
}
}
func (s *S) TestReadGFF(c *check.C) {
for i, g := range []struct {
gff string
feat []*gff.Feature
}{
{
gff: `SEQ1 EMBL atg 103 105 . + 0
SEQ1 EMBL exon 103 172 . + 0
SEQ1 EMBL splice5 172 173 . + .
SEQ1 netgene splice5 172 173 0.94 + .
SEQ1 genie sp5-20 163 182 2.3 + .
SEQ1 genie sp5-10 168 177 2.1 + .
SEQ2 grail ATG 17 19 2.1 - 0
`,
feat: []*gff.Feature{
{SeqName: "SEQ1", Source: "EMBL", Feature: "atg", FeatStart: 102, FeatEnd: 105, FeatScore: nil, FeatFrame: gff.Frame0, FeatStrand: seq.Plus},
{SeqName: "SEQ1", Source: "EMBL", Feature: "exon", FeatStart: 102, FeatEnd: 172, FeatScore: nil, FeatFrame: gff.Frame0, FeatStrand: seq.Plus},
{SeqName: "SEQ1", Source: "EMBL", Feature: "splice5", FeatStart: 171, FeatEnd: 173, FeatScore: nil, FeatFrame: gff.NoFrame, FeatStrand: seq.Plus},
{SeqName: "SEQ1", Source: "netgene", Feature: "splice5", FeatStart: 171, FeatEnd: 173, FeatScore: floatPtr(0.94), FeatFrame: gff.NoFrame, FeatStrand: seq.Plus},
{SeqName: "SEQ1", Source: "genie", Feature: "sp5-20", FeatStart: 162, FeatEnd: 182, FeatScore: floatPtr(2.3), FeatFrame: gff.NoFrame, FeatStrand: seq.Plus},
{SeqName: "SEQ1", Source: "genie", Feature: "sp5-10", FeatStart: 167, FeatEnd: 177, FeatScore: floatPtr(2.1), FeatFrame: gff.NoFrame, FeatStrand: seq.Plus},
{SeqName: "SEQ2", Source: "grail", Feature: "ATG", FeatStart: 16, FeatEnd: 19, FeatScore: floatPtr(2.1), FeatFrame: gff.Frame0, FeatStrand: seq.Minus},
},
},
} {
sc := featio.NewScanner(
gff.NewReader(
bytes.NewBufferString(g.gff),
),
)
var j int
for sc.Next() {
f := sc.Feat()
c.Check(f, check.DeepEquals, g.feat[j], check.Commentf("Test: %d Line: %d", i, j+1))
j++
}
c.Check(sc.Error(), check.Equals, nil)
c.Check(j, check.Equals, len(g.feat))
}
}
func readMappings(file string) (map[string]*gff.Feature, error) {
f, err := os.Open(file)
if err != nil {
return nil, err
}
mapping := make(map[string]*gff.Feature)
sc := featio.NewScanner(gff.NewReader(f))
for id := uintptr(1); sc.Next(); id++ {
f := sc.Feat().(*gff.Feature)
read := f.FeatAttributes.Get("Read")
if read == "" {
continue
}
// Currently reefer only expects a single hit per read.
mapping[strings.Fields(read)[0]] = f
}
if err != nil {
log.Fatalf("error during GFF read: %v", err)
}
return mapping, nil
}
func readBED(in string) ([]*bed.Bed3, error) {
bf, err := os.Open(in)
if err != nil {
return nil, err
}
defer bf.Close()
br, err := bed.NewReader(bf, 3)
if err != nil {
return nil, err
}
var fs []*bed.Bed3
sc := featio.NewScanner(br)
for sc.Next() {
fs = append(fs, sc.Feat().(*bed.Bed3))
}
err = sc.Error()
if err != nil {
return nil, err
}
return fs, nil
}
func main() {
groups := make(map[string]struct {
chrom string
start, end int
})
sc := featio.NewScanner(gff.NewReader(os.Stdin))
for sc.Next() {
f := sc.Feat().(*gff.Feature)
g := f.FeatAttributes.Get("Group")
if g == "" {
continue
}
grp, ok := groups[g]
if !ok {
groups[g] = struct {
chrom string
start, end int
}{chrom: f.SeqName, start: f.FeatStart, end: f.FeatEnd}
continue
}
if f.FeatStart < grp.start {
grp.start = f.FeatStart
}
if grp.end < f.FeatEnd {
grp.end = f.FeatEnd
}
groups[g] = grp
}
if err := sc.Error(); err != nil {
log.Fatalf("error during gff read: %v", err)
}
for k, v := range groups {
fmt.Printf("%s\t%d\t%d\t%s\n", v.chrom, v.start, v.end, k)
}
}
func main() {
vm := make(map[string]*vector)
fm := make(map[string]string)
sc := featio.NewScanner(gff.NewReader(os.Stdin))
for sc.Next() {
f := sc.Feat().(*gff.Feature)
if f.Feature != "repeat" {
continue
}
att := f.FeatAttributes.Get("repeat")
if att == "" {
continue
}
fields := strings.Fields(att)
s := mustAtoi(fields[2])
e := mustAtoi(fields[3])
v, ok := vm[fields[0]]
if !ok {
v = &vector{}
vm[fields[0]] = v
fm[fields[0]] = fields[1]
}
for i := e; i >= s; i-- {
v.inc(i)
}
}
for typ, vec := range vm {
for pos, val := range *vec {
if val != 0 {
fmt.Printf("%s\t%s\t%d\t%d\n", fm[typ], typ, pos, val)
}
}
}
}
func main() {
nameSet := make(map[string]struct{})
sc := featio.NewScanner(gff.NewReader(os.Stdin))
for sc.Next() {
f := sc.Feat().(*gff.Feature)
n := f.FeatAttributes.Get("Read")
if n == "" {
continue
}
nameSet[n] = struct{}{}
}
if err := sc.Error(); err != nil {
log.Fatalf("error during gff read: %v", err)
}
names := make([]string, 0, len(nameSet))
for n := range nameSet {
names = append(names, n)
}
sort.Strings(names)
for _, n := range names {
fmt.Println(n)
}
}
func main() {
if len(os.Args) < 2 {
fmt.Fprintln(os.Stderr, "invalid invocation: must have at least one reads file")
os.Exit(1)
}
extract := make(map[string][2]int)
sc := featio.NewScanner(gff.NewReader(os.Stdin))
for sc.Next() {
f := sc.Feat().(*gff.Feature)
read := f.FeatAttributes.Get("Read")
if read == "" {
continue
}
fields := strings.Fields(read)
name := fields[0]
start, err := strconv.Atoi(fields[1])
if err != nil {
log.Fatalf("failed to parse %q: %v", read, err)
}
end, err := strconv.Atoi(fields[2])
if err != nil {
log.Fatalf("failed to parse %q: %v", read, err)
}
extract[name] = [2]int{start, end}
}
err := sc.Error()
if err != nil {
log.Fatalf("error during GFF read: %v", err)
}
for _, reads := range os.Args[1:] {
sf, err := os.Open(reads)
if err != nil {
log.Fatalf("failed to open %q: %v", reads, err)
}
sr, err := sam.NewReader(sf)
if err != nil {
log.Fatalf("failed to open SAM input %q: %v", reads, err)
}
for {
r, err := sr.Read()
if err != nil {
if err != io.EOF {
log.Fatalf("unexpected error reading SAM: %v", err)
}
break
}
v, ok := extract[r.Name]
if !ok {
continue
}
// Currently reefer only expects a single hit per read,
// so any multiples are due to duplicate read file input.
// Update this behaviour if we change reefer to look at
// remapping soft-clipped segments.
delete(extract, r.Name)
reverse := r.Flags&sam.Reverse != 0
rng := fmt.Sprintf("//%d_%d", v[0], v[1])
if reverse {
rng += "(-)"
len := r.Seq.Length
v[0], v[1] = len-v[1], len-v[0]
}
v[0] = feat.OneToZero(v[0])
s := linear.NewSeq(
r.Name+rng,
alphabet.BytesToLetters(r.Seq.Expand())[v[0]:v[1]],
alphabet.DNA,
)
if reverse {
s.Desc = "(sequence revcomp relative to read)"
}
fmt.Printf("%60a\n", s)
}
sf.Close()
}
}
func main() {
flag.Parse()
if *in == "" || *ref == "" || *mapfile == "" || *contigs == "" {
flag.Usage()
os.Exit(0)
}
refTrees, err := readAnnotations(*ref)
if err != nil {
log.Fatalf("failed to read annotation trees: %v", err)
}
mapping, err := readMappings(*mapfile)
if err != nil {
log.Fatalf("failed to read mapping file: %v", err)
}
contigLength, err := readContigs(*contigs)
if err != nil {
log.Fatalf("failed to read contig file: %v", err)
}
f, err := os.Open(*in)
if err != nil {
log.Fatalf("failed to open %q: %v", *in, err)
}
w := gff.NewWriter(os.Stdout, 60, true)
sc := featio.NewScanner(gff.NewReader(f))
for sc.Next() {
f := sc.Feat().(*gff.Feature)
ok, err := within(*buf, f.SeqName)
if err != nil {
log.Fatalf("failed to parse sequence name: %s: %v", f.SeqName, err)
}
if !ok {
log.Printf("too close to read end: excluding %+v", f)
continue
}
repeat := f.FeatAttributes.Get("Repeat")
if repeat == "" {
continue
}
fields := strings.Fields(repeat)
name := strings.Split(f.SeqName, "//")
if len(name) != 2 {
log.Fatalf("unexpected sequence name in input: %q", f.SeqName)
}
contigSide, ok := mapping[name[0]]
if !ok {
log.Fatalf("unexpected sequence name in input: %q", f.SeqName)
}
if contigSide.FeatStart+f.FeatStart < *buf {
log.Printf("too close to contig start:\n\texcluding %#v\n\tcontig %#v\n\n%d < %d", f, contigSide, contigSide.FeatStart, *buf)
continue
}
length, ok := contigLength[contigSide.SeqName]
if !ok {
log.Fatalf("unexpected sequence name in contig mapping: %q", contigSide.SeqName)
}
if length-((contigSide.FeatEnd-contigSide.FeatStart)+f.FeatEnd) < *buf {
log.Printf("too close to contig end:\n\texcluding %#v\n\tcontig %#v", f, contigSide)
continue
}
t, ok := refTrees[contigSide.SeqName]
if !ok {
log.Fatalf("no tree for %v mapped by %v", contigSide.SeqName, f.SeqName)
}
var n int
hits := t.Get(gffInterval{Feature: contigSide})
for _, h := range hits {
f := h.(gffInterval)
repeat := f.FeatAttributes.Get("Repeat")
if repeat == "" {
continue
}
hitClass := strings.Fields(repeat)[1]
if fields[1] == hitClass {
n++
}
}
if n != 0 {
log.Printf("too many hits: excluding %+v", f)
for _, h := range hits {
log.Printf("\t%+v", h.(gffInterval).Feature)
}
continue
}
w.Write(f)
}
err = sc.Error()
if err != nil {
log.Fatalf("error during GFF read: %v", err)
}
}
func main() {
flag.Parse()
if *in == "" {
flag.Usage()
os.Exit(1)
}
f, err := os.Open(*in)
if err != nil {
log.Fatalf("failed to open %q: %v", *in, err)
}
defer f.Close()
names := make(map[string]map[string]struct{})
sc := featio.NewScanner(gff.NewReader(f))
for sc.Next() {
feat := sc.Feat().(*gff.Feature)
read := feat.FeatAttributes.Get("Read")
if read == "" {
continue
}
read = strings.Fields(read)[0]
idx := strings.LastIndex(read, "/")
e, ok := names[read[:idx]]
if !ok {
e = make(map[string]struct{})
names[read[:idx]] = e
}
e[read[idx+1:]] = struct{}{}
}
if err := sc.Error(); err != nil {
log.Fatalf("error during fasta read: %v", err)
}
f.Close()
base := filepath.Base(*in)
unique, err := os.Create(base + ".unique.text")
if err != nil {
log.Fatalf("failed to create %q: %v", base+".unique.text", err)
}
defer unique.Close()
nonUnique, err := os.Create(base + ".non-unique.text")
if err != nil {
log.Fatalf("failed to create %q: %v", base+".non-unique.text", err)
}
defer nonUnique.Close()
for name, coords := range names {
switch len(coords) {
case 0:
case 1:
fmt.Fprintln(unique, name)
default:
s := make([]string, 0, len(coords))
for c := range coords {
s = append(s, c)
}
sort.Strings(s)
fmt.Fprintf(nonUnique, "%s\t%v\n", name, s)
}
}
}
func main() {
flag.Var(&alnmat, "align", "specify the match, mismatch and gap parameters")
flag.Parse()
if *in == "" {
flag.Usage()
os.Exit(1)
}
f, err := os.Open(*in)
if err != nil {
log.Fatalf("failed to open %q: %v", *in, err)
}
events := make(map[string][]*gff.Feature)
fsc := featio.NewScanner(gff.NewReader(f))
for fsc.Next() {
f := fsc.Feat().(*gff.Feature)
fields := strings.Fields(f.FeatAttributes.Get("Read"))
if len(fields) != 3 {
log.Fatalf("bad record: %+v", f)
}
events[fields[0]] = append(events[fields[0]], f)
}
if err := fsc.Error(); err != nil {
log.Fatalf("error during gff read: %v", err)
}
f.Close()
w := gff.NewWriter(os.Stdout, 60, true)
w.WriteComment("Right coordinates (field 5) and strand (field 7) are hypothetical.")
var out *os.File
if *fastaOut != "" {
out, err = os.Create(*fastaOut)
if err != nil {
log.Fatalf("failed to create fasta insertion output file %q: %v", *fastaOut, err)
}
defer out.Close()
}
hw := *window / 2
sw := makeTable(alphabet.DNAgapped, alnmat)
for _, ref := range flag.Args() {
f, err = os.Open(ref)
if err != nil {
log.Fatalf("failed to open reference %q: %v", ref, err)
}
ssc := seqio.NewScanner(fasta.NewReader(f, linear.NewSeq("", nil, alphabet.DNAgapped)))
loop:
for ssc.Next() {
seq := ssc.Seq().(*linear.Seq)
for _, f := range events[seq.Name()] {
fields := strings.Fields(f.FeatAttributes.Get("Read"))
if len(fields) != 3 {
log.Fatalf("bad record: %+v", f)
}
start, err := strconv.Atoi(fields[1])
if err != nil {
log.Fatalf("failed to get start coordinate: %v", err)
}
end, err := strconv.Atoi(fields[2])
if err != nil {
log.Fatalf("failed to get end coordinate: %v", err)
}
if out != nil {
insert := *seq
if insert.Desc != "" {
insert.Desc += " "
}
insert.Desc += fmt.Sprintf("[%d,%d)", start, end)
insert.Seq = insert.Seq[start:end]
fmt.Fprintf(out, "%60a\n", &insert)
}
var lOff, lEnd, rOff, rEnd int
// If we have refined ends, use them.
if dup := f.FeatAttributes.Get("Dup"); dup != "" {
d, err := strconv.Atoi(dup)
if err != nil {
log.Fatalf("failed to get duplication length: %v", err)
}
lOff = max(0, start-d)
lEnd = start
rOff = end
rEnd = min(len(seq.Seq), end+d)
} else {
lOff = max(0, start-hw)
lEnd = min(len(seq.Seq), start+hw)
rOff = max(0, end-hw)
rEnd = min(len(seq.Seq), end+hw)
// Ensure windows don't overlap.
if lEnd > rOff {
lEnd = (lEnd + rOff) / 2
rOff = lEnd
}
}
if lEnd-lOff < *thresh || rEnd-rOff < *thresh {
// Don't do fruitless work.
continue loop
}
left := *seq
left.ID = "prefix"
left.Seq = left.Seq[lOff:lEnd]
right := *seq
right.ID = "postfix"
right.Seq = right.Seq[rOff:rEnd]
aln, err := sw.Align(&right, &left)
if err != nil {
log.Fatal(err)
}
fa := align.Format(&right, &left, aln, '-')
for _, seg := range fa {
var n int
for _, l := range seg.(alphabet.Letters) {
if l != '-' {
n++
}
}
if n < *thresh {
continue loop
}
}
var sc int
for _, seg := range aln {
type scorer interface {
Score() int
}
sc += seg.(scorer).Score()
}
f.FeatAttributes = append(f.FeatAttributes, gff.Attribute{
Tag: "TSD", Value: fmt.Sprintf(`%v %d %d %v "%v" %d`,
fa[0], aln[len(aln)-1].Features()[0].End()+lOff,
aln[0].Features()[1].Start()+rOff, fa[1],
aln, sc),
})
w.Write(f)
}
}
if err := ssc.Error(); err != nil {
log.Fatalf("error during fasta read: %v", err)
}
f.Close()
}
}