1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387
|
// Copyright ©2011-2013 The bíogo Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Package pals implements functions and methods required for PALS sequence alignment.
package pals
import (
"github.com/biogo/biogo/align/pals/dp"
"github.com/biogo/biogo/align/pals/filter"
"github.com/biogo/biogo/index/kmerindex"
"github.com/biogo/biogo/morass"
"github.com/biogo/biogo/seq/linear"
"github.com/biogo/biogo/util"
"errors"
"io"
"unsafe"
)
// Default thresholds for filter and alignment.
var (
DefaultLength = 400
DefaultMinIdentity = 0.94
MaxAvgIndexListLen = 15.0
TubeOffsetDelta = 32
)
// Default filter and dynamic programming Cost values.
const (
MaxIGap = 5
DiffCost = 3
SameCost = 1
MatchCost = DiffCost + SameCost
BlockCost = DiffCost * MaxIGap
RMatchCost = float64(DiffCost) + 1
)
// A dp.Costs based on the default cost values.
var defaultCosts = dp.Costs{
MaxIGap: MaxIGap,
DiffCost: DiffCost,
SameCost: SameCost,
MatchCost: MatchCost,
BlockCost: BlockCost,
RMatchCost: RMatchCost,
}
// Default word characteristics.
var (
MinWordLength = 4 // For minimum word length, choose k=4 arbitrarily.
MaxKmerLen = 15 // Currently limited to 15 due to 32 bit int limit for indexing slices
)
// PALS is a type that can perform pairwise alignments of large sequences based on the papers:
// PILER: identification and classification of genomic repeats.
// Robert C. Edgar and Eugene W. Myers. Bioinformatics Suppl. 1:i152-i158 (2005)
// Efficient q-gram filters for finding all 𝛜-matches over a given length.
// Kim R. Rasmussen, Jens Stoye, and Eugene W. Myers. J. of Computational Biology 13:296–308 (2006).
type PALS struct {
target, query *linear.Seq
selfCompare bool
index *kmerindex.Index
FilterParams *filter.Params
DPParams *dp.Params
dp.Costs
log Logger
timer *util.Timer
tubeOffset int
maxMem *uintptr
hitFilter *filter.Filter
morass *morass.Morass
trapezoids filter.Trapezoids
err error
threads int
}
// Return a new PALS aligner. Requires
func New(target, query *linear.Seq, selfComp bool, m *morass.Morass, tubeOffset int, mem *uintptr, log Logger) *PALS {
return &PALS{
target: target,
query: query,
selfCompare: selfComp,
log: log,
tubeOffset: tubeOffset,
Costs: defaultCosts,
maxMem: mem,
morass: m,
}
}
// Optimise the PALS parameters for given memory, kmer length, hit length and sequence identity.
// An error is returned if no satisfactory parameters can be found.
func (p *PALS) Optimise(minHitLen int, minId float64) error {
if minId < 0 || minId > 1.0 {
return errors.New("pals: minimum identity out of range")
}
if minHitLen <= MinWordLength {
return errors.New("pals: minimum hit length too short")
}
if p.log != nil {
p.log.Print("Optimising filter parameters")
}
filterParams := &filter.Params{}
// Lower bound on word length k by requiring manageable index.
// Given kmer occurs once every 4^k positions.
// Hence average number of index entries is i = N/(4^k) for random
// string of length N.
// Require i <= I, then k > log_4(N/i).
minWordSize := int(util.Log4(float64(p.target.Len())) - util.Log4(MaxAvgIndexListLen) + 0.5)
// First choice is that filter criteria are same as DP criteria,
// but this may not be possible.
seedLength := minHitLen
seedDiffs := int(float64(minHitLen) * (1 - minId))
// Find filter valid filter parameters, starting from preferred case.
for {
minWords := -1
if MaxKmerLen < minWordSize {
if p.log != nil {
p.log.Printf("Word size too small: %d < %d\n", MaxKmerLen, minWordSize)
}
}
for wordSize := MaxKmerLen; wordSize >= minWordSize; wordSize-- {
filterParams.WordSize = wordSize
filterParams.MinMatch = seedLength
filterParams.MaxError = seedDiffs
if p.tubeOffset > 0 {
filterParams.TubeOffset = p.tubeOffset
} else {
filterParams.TubeOffset = filterParams.MaxError + TubeOffsetDelta
}
mem := p.MemRequired(filterParams)
if p.maxMem != nil && mem > *p.maxMem {
if p.log != nil {
p.log.Printf("Parameters n=%d k=%d e=%d, mem=%d MB > maxmem=%d MB\n",
filterParams.MinMatch,
filterParams.WordSize,
filterParams.MaxError,
mem/1e6,
*p.maxMem/1e6)
}
minWords = -1
continue
}
minWords = filter.MinWordsPerFilterHit(seedLength, wordSize, seedDiffs)
if minWords <= 0 {
if p.log != nil {
p.log.Printf("Parameters n=%d k=%d e=%d, B=%d\n",
filterParams.MinMatch,
filterParams.WordSize,
filterParams.MaxError,
minWords)
}
minWords = -1
continue
}
length := p.AvgIndexListLength(filterParams)
if length > MaxAvgIndexListLen {
if p.log != nil {
p.log.Printf("Parameters n=%d k=%d e=%d, B=%d avgixlen=%.2f > max = %.2f\n",
filterParams.MinMatch,
filterParams.WordSize,
filterParams.MaxError,
minWords,
length,
MaxAvgIndexListLen)
}
minWords = -1
continue
}
break
}
if minWords > 0 {
break
}
// Failed to find filter parameters, try
// fewer errors and shorter seed.
if seedLength >= minHitLen/4 {
seedLength /= 2
continue
}
if seedDiffs > 0 {
seedDiffs--
continue
}
return errors.New("pals: failed to find filter parameters")
}
p.FilterParams = filterParams
p.DPParams = &dp.Params{
MinHitLength: minHitLen,
MinId: minId,
}
return nil
}
// Return an estimate of the average number of hits for any given kmer.
func (p *PALS) AvgIndexListLength(filterParams *filter.Params) float64 {
d := int(1) << (uint(filterParams.WordSize) * 2)
return float64(p.target.Len()) / float64(d)
}
// Return an estimate of the amount of memory required for the filter.
func (p *PALS) filterMemRequired(filterParams *filter.Params) uintptr {
words := util.Pow4(filterParams.WordSize)
tubeWidth := filterParams.TubeOffset + filterParams.MaxError
maxActiveTubes := (p.target.Len()+tubeWidth-1)/filterParams.TubeOffset + 1
tubes := uintptr(maxActiveTubes) * unsafe.Sizeof(tubeState{})
finger := unsafe.Sizeof(uint32(0)) * uintptr(words)
pos := unsafe.Sizeof(0) * uintptr(p.target.Len())
return finger + pos + tubes
}
// filter.tubeState is repeated here to allow memory calculation without
// exporting tubeState from filter package.
type tubeState struct {
QLo int
QHi int
Count int
}
// Return an estimate of the total amount of memory required.
func (p *PALS) MemRequired(filterParams *filter.Params) uintptr {
filter := p.filterMemRequired(filterParams)
sequence := uintptr(p.target.Len()) + unsafe.Sizeof(p.target)
if p.target != p.query {
sequence += uintptr(p.query.Len()) + unsafe.Sizeof(p.query)
}
return filter + sequence
}
// Build the kmerindex for filtering.
func (p *PALS) BuildIndex() error {
p.notify("Indexing")
ki, err := kmerindex.New(p.FilterParams.WordSize, p.target)
if err != nil {
return err
} else {
ki.Build()
p.notify("Indexed")
}
p.index = ki
p.hitFilter = filter.New(p.index, p.FilterParams)
return nil
}
// Share allows the receiver to use the index and parameters of m.
func (p *PALS) Share(m *PALS) {
p.index = m.index
p.FilterParams = m.FilterParams
p.DPParams = m.DPParams
p.hitFilter = filter.New(p.index, p.FilterParams)
}
// Align performs filtering and alignment for one strand of query.
func (p *PALS) Align(complement bool) (dp.Hits, error) {
if p.err != nil {
return nil, p.err
}
var (
working *linear.Seq
err error
)
if complement {
p.notify("Complementing query")
working = p.query.Clone().(*linear.Seq)
working.RevComp()
p.notify("Complemented query")
} else {
working = p.query
}
p.notify("Filtering")
err = p.hitFilter.Filter(working, p.selfCompare, complement, p.morass)
if err != nil {
return nil, err
}
p.notifyf("Identified %d filter hits", p.morass.Len())
p.notify("Merging")
merger := filter.NewMerger(p.index, working, p.FilterParams, p.MaxIGap, p.selfCompare)
var h filter.Hit
for {
if err = p.morass.Pull(&h); err != nil {
break
}
merger.MergeFilterHit(&h)
}
if err != nil && err != io.EOF {
return nil, err
}
p.err = p.morass.Clear()
p.trapezoids = merger.FinaliseMerge()
lt, lq := p.trapezoids.Sum()
p.notifyf("Merged %d trapezoids covering %d x %d", len(p.trapezoids), lt, lq)
p.notify("Aligning")
aligner := dp.NewAligner(
p.target, working,
p.FilterParams.WordSize, p.DPParams.MinHitLength, p.DPParams.MinId,
)
aligner.Costs = &p.Costs
hits := aligner.AlignTraps(p.trapezoids)
hitCoverageA, hitCoverageB, err := hits.Sum()
if err != nil {
return nil, err
}
p.notifyf("Aligned %d hits covering %d x %d", len(hits), hitCoverageA, hitCoverageB)
return hits, nil
}
// Trapezoids returns the filter trapezoids identified during a call to Align.
func (p *PALS) Trapezoids() filter.Trapezoids { return p.trapezoids }
// AlignFrom performs filtering and alignment for one strand of query using the
// provided filter trapezoids as seeds.
func (p *PALS) AlignFrom(traps filter.Trapezoids, complement bool) (dp.Hits, error) {
if p.err != nil {
return nil, p.err
}
var (
working *linear.Seq
err error
)
if complement {
p.notify("Complementing query")
working = p.query.Clone().(*linear.Seq)
working.RevComp()
p.notify("Complemented query")
} else {
working = p.query
}
p.notify("Aligning")
aligner := dp.NewAligner(
p.target, working,
p.FilterParams.WordSize, p.DPParams.MinHitLength, p.DPParams.MinId,
)
aligner.Costs = &p.Costs
hits := aligner.AlignTraps(traps)
hitCoverageA, hitCoverageB, err := hits.Sum()
if err != nil {
return nil, err
}
p.notifyf("Aligned %d hits covering %d x %d", len(hits), hitCoverageA, hitCoverageB)
return hits, nil
}
// Remove file system components of filter. This should be called after
// the last use of the aligner.
func (p *PALS) CleanUp() error { return p.morass.CleanUp() }
// Interface for logger used by PALS.
type Logger interface {
Print(v ...interface{})
Printf(format string, v ...interface{})
}
func (p *PALS) notify(n string) {
if p.log != nil {
p.log.Print(n)
}
}
func (p *PALS) notifyf(f string, n ...interface{}) {
if p.log != nil {
p.log.Printf(f, n...)
}
}
|