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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453
|
package structures;
import java.util.ArrayList;
import java.util.concurrent.ArrayBlockingQueue;
import java.util.concurrent.locks.ReadWriteLock;
import java.util.concurrent.locks.ReentrantReadWriteLock;
import dna.Scaffold;
import shared.Timer;
import shared.Tools;
import template.Accumulator;
import template.ThreadWaiter;
public class StandardDeviator implements Accumulator<StandardDeviator.ProcessThread> {
public StandardDeviator(boolean stranded_, int strand_) {
stranded=stranded_;
strand=strand_;
}
//This one is to slow, maybe because short contigs need too much locking.
//They would have to be put into lists of contigs.
// public boolean calculateStuffQueued(int threads0, Collection<Scaffold> scaffolds,
// boolean calcStd, boolean calcMedian, double windowAvg, int window) {
//
// final int threads=Tools.mid(1, threads0, scaffolds.size());
// Timer t=new Timer();
//
// System.err.println("threads="+threads);
//
// {//First pass to sum depth
// totalSum=0;
// bins=0;
// ArrayBlockingQueue<Scaffold> queue=new ArrayBlockingQueue<Scaffold>(threads*2);
// //Fill a list with ProcessThreads
// ArrayList<ProcessThread> alpt=new ArrayList<ProcessThread>(threads);
// for(int i=0; i<threads; i++){
// alpt.add(new ProcessThread(i, queue, false, false, -1, -1, -1));
// }//TODO: For some reason this is running singlethreaded... 97% CPU.
// //Start the threads and wait for them to finish
// ThreadWaiter.startThreads(alpt);
// for(Scaffold s : scaffolds) {addToQueue(queue, s);}
// addToQueue(queue, POISON);
// boolean success=ThreadWaiter.waitForThreads(alpt, this);
// errorState&=!success;
// t.stopAndStart("calcSumMT:");
// }
//
// final double globalMean=totalSum/(double)bins;
// {//Second pass for statistics
// totalSum=0;
// bins=0;
// ArrayBlockingQueue<Scaffold> queue=new ArrayBlockingQueue<Scaffold>(threads*2);
// //Fill a list with ProcessThreads
// ArrayList<ProcessThread> alpt=new ArrayList<ProcessThread>(threads);
// for(int i=0; i<threads; i++){
// alpt.add(new ProcessThread(i, queue, calcStd, calcMedian, globalMean, windowAvg, window));
// }//This part only uses 150% CPU
// //Start the threads and wait for them to finish
// ThreadWaiter.startThreads(alpt);
// for(Scaffold s : scaffolds) {addToQueue(queue, s);}
// addToQueue(queue, POISON);
// boolean success=ThreadWaiter.waitForThreads(alpt, this);
// errorState&=!success;
// t.stopAndStart("calcStdMT:");
// }
//
// return errorState;
// }
public boolean calculateStuff(int threads0, ArrayList<Scaffold> scaffolds,
boolean calcStd, boolean calcMedian, boolean calcHist,
int minDepth, int histMax, double windowAvg, int window) {
final int threads=Tools.mid(1, threads0, scaffolds.size());
Timer t=new Timer();
// if(verbose) {System.err.println("threads="+threads);}
{//First pass to sum depth for global stdev
totalSum=0;
bins=0;
//Fill a list with ProcessThreads
ArrayList<ProcessThread> alpt=new ArrayList<ProcessThread>(threads);
for(int i=0; i<threads; i++){
alpt.add(new ProcessThread(i, threads, scaffolds, false, false, false, minDepth, -1, -1, -1, -1));
}
//Start the threads and wait for them to finish
ThreadWaiter.startThreads(alpt);
boolean success=ThreadWaiter.waitForThreadsToFinish(alpt, this);
errorState&=!success;
if(verbose) {t.stopAndStart("calcSumMT:");}
}
final double globalMean=totalSum/(double)bins;
{//Second pass for statistics
totalSum=0;
bins=0;
//Fill a list with ProcessThreads
ArrayList<ProcessThread> alpt=new ArrayList<ProcessThread>(threads);
for(int i=0; i<threads; i++){
alpt.add(new ProcessThread(i, threads, scaffolds,
calcStd, calcMedian, calcHist, minDepth, histMax, globalMean, windowAvg, window));
}
//Start the threads and wait for them to finish
ThreadWaiter.startThreads(alpt);
boolean success=ThreadWaiter.waitForThreadsToFinish(alpt, this);
errorState&=!success;
if(verbose) {t.stopAndStart("calcStdMT:");}
}
return errorState;
}
public double[] standardDeviation(ArrayList<Scaffold> scaffolds, int minscaf0){
totalSum=bins=0;
mean=stdev=0;
long scafsCounted=0;
final int minscaf=Tools.max(minscaf0, 1);
for(Scaffold scaf : scaffolds){
if(scaf.length>=minscaf){
final CoverageArray ca=(CoverageArray)(strand==1 ? scaf.obj1 : scaf.obj0);
bins+=scaf.length;
totalSum+=(ca==null ? 0 : ca.sum());
scafsCounted++;
}
}
if(bins<1){return new double[] {mean, stdev};}
mean=totalSum/(double)bins;
// System.err.println("sum="+totalSum+", bins="+bins+", mean="+mean+", scaffolds="+scafsCounted);
double sumdev2=0;
for(Scaffold scaf : scaffolds){
if(scaf.length>=minscaf){
final CoverageArray ca=(CoverageArray)(strand==1 ? scaf.obj1 : scaf.obj0);
if(ca==null) {
sumdev2+=(mean*mean*scaf.length);
}else {
sumdev2+=ca.devSum(mean);
}
}
}
stdev=Math.sqrt(sumdev2/bins);
// System.err.println("sumdev2="+sumdev2+", stdev="+stdev);
return new double[] {mean, stdev};
}
/** This works but is slow. Just use unbinned std. */
@Deprecated
public double[] standardDeviationBinned(ArrayList<Scaffold> scaffolds, int binsize, int minscaf0){
totalSumBinned=binsBinned=0;
meanBinned=stdevBinned=0;
final int minscaf=Tools.max(minscaf0, 1);
for(Scaffold scaf : scaffolds){
if(scaf.length>=minscaf){
CoverageArray ca=(CoverageArray)(strand==1 ? scaf.obj1 : scaf.obj0);
int bins=(int)((scaf.length+(long)binsize-1L)/binsize);
assert(bins>0) : scaf.length+", "+binsize+", "+minscaf;
binsBinned+=bins;
if(ca==null) {
totalSumBinned+=0;
}else {
totalSumBinned+=ca.sum()/bins;
}
}
}
if(binsBinned<1){return new double[] {meanBinned, stdevBinned};}
meanBinned=totalSumBinned/(double)binsBinned;
double sumdev2=0;
for(Scaffold scaf : scaffolds){
if(scaf.length>=minscaf){
CoverageArray ca=(CoverageArray)(strand==1 ? scaf.obj1 : scaf.obj0);
int lastPos=-1, nextPos=binsize-1;
long tempSum=0;
final int lim=scaf.length-1;
for(int i=0; i<scaf.length; i++){
int x=(ca==null ? 0 : ca.get(i));
tempSum+=x;
if(i>=nextPos || i==lim){
int bin=(i-lastPos);
double depth=(tempSum/(double)bin);
double dev=meanBinned-depth;
sumdev2+=(dev*dev);
nextPos+=binsize;
lastPos=i;
tempSum=0;
}
}
}
}
stdevBinned=Math.sqrt(sumdev2/binsBinned);
return new double[] {meanBinned, stdevBinned};
}
//Unsafe because it will fail if there are over 2 billion bins
public double[] standardDeviationBinnedUnsafe(String fname, ArrayList<Scaffold> scaffolds, int binsize, int minscaf0){
final int minscaf=Tools.max(minscaf0, 1);
LongList list=new LongList();
for(Scaffold scaf : scaffolds){
CoverageArray ca=(CoverageArray)(strand==1 ? scaf.obj1 : scaf.obj0);
int lastPos=-1, nextPos=binsize-1;
long sum=0;
final int lim=scaf.length-1;
for(int i=0; i<scaf.length; i++){
int x=(ca==null ? 0 : ca.get(i));
sum+=x;
if(i>=nextPos || i==lim){
int bin=(i-lastPos);
if(scaf.length>=minscaf){
list.add((int)(10*(sum/(double)bin)));
}
nextPos+=binsize;
lastPos=i;
sum=0;
}
}
}
list.sort();
mean=0.1*list.mean();
double median=0.1*list.median();
double mode=0.1*list.mode();
stdev=0.1*list.stdev();
return new double[] {mean, median, mode, stdev};
}
/*--------------------------------------------------------------*/
/*---------------- Threading ----------------*/
/*--------------------------------------------------------------*/
// static class ProcessThreadQueued extends Thread {
//
// ProcessThreadQueued(int tid_, ArrayBlockingQueue<Scaffold> queue_, boolean calcStd_,
// boolean calcMedian_, double globalMean_, double windowAvg_, int window_){
// tid=tid_;
// queue=queue_;
//
// calcStd=calcStd_;
// calcMedian=calcMedian_;
// globalMean=globalMean_;
// windowAvg=windowAvg_;
// window=window_;
// }
//
// @Override
// public void run() {
//// synchronized(this) {
// for(Scaffold next=getNext(queue); next!=POISON; next=getNext(queue)){
//// synchronized(next) {
// binsT+=next.length;
// processCA((CoverageArray)next.obj1);
// processCA((CoverageArray)next.obj2);
//// }
// }
// addToQueue(queue, POISON);//Send it to the next one
//// }
// }
//
// private void processCA(CoverageArray ca) {
// if(ca==null) {return;}
// totalSumT+=ca.sum();
// if(calcStd) {ca.standardDeviation();}
// if(calcMedian) {ca.median();}
// if(globalMean>-1) {ca.devSum(globalMean);}
// if(windowAvg>0) {ca.basesUnderAverageCoverage(windowAvg, window);}
// }
//
// long totalSumT=0;
// long binsT=0;
//
// final int tid;
// final ArrayBlockingQueue<Scaffold> queue;
//
// final boolean calcStd;
// final boolean calcMedian;
// final double globalMean;
// final double windowAvg;
// final int window;
// }
class ProcessThread extends Thread {
ProcessThread(int tid_, int threads_, ArrayList<Scaffold> list_,
boolean calcStd_, boolean calcMedian_, boolean calcHist_, int minDepth_, int histMax_,
double globalMean_, double windowAvg_, int window_){
// System.err.println(calcHist_);
tid=tid_;
threads=threads_;
list=list_;
calcStd=calcStd_;
calcMedian=calcMedian_;
calcHist=calcHist_;
minDepth=minDepth_;
histMax=histMax_;
globalMean=globalMean_;
windowAvg=windowAvg_;
window=window_;
if(calcHist) {
assert(histMax>0) : histMax;
hist0=new LongList(Tools.min(histMax+1, Character.MAX_VALUE+1));
if(stranded) {hist1=new LongList(Tools.min(histMax+1, Character.MAX_VALUE+1));}
}
}
@Override
public synchronized void run() {
for(int i=tid; i<list.size(); i+=threads){
Scaffold scaf=list.get(i);
synchronized(scaf) {
binsT+=scaf.length;
processCA((CoverageArray)scaf.obj0, 0, scaf.length, scaf.basehits);
if(stranded) {processCA((CoverageArray)scaf.obj1, 1, scaf.length, scaf.basehits);}
}
}
}
/** In the case of preallocated ca's with 0 mapped reads, this goes faster */
private void processCA(CoverageArray ca, int strand, int length, long mappedBases) {
if(mappedBases>0) {
if(calcHist) {addToHist(ca, strand, length);}
if(ca==null) {return;}
totalSumT+=ca.calcSumAndCovered(minDepth);
if(calcStd) {ca.standardDeviation();}
if(calcMedian) {ca.median();}
if(globalMean>-1) {ca.devSum(globalMean);}
if(windowAvg>0) {ca.basesUnderAverageCoverage(windowAvg, window);}
}else {
if(calcHist) {addToHist(null, strand, length);}
if(ca==null) {return;}
totalSumT+=0;
ca.setSum(0);
ca.setCovered(0);
ca.setStdev(0);
ca.setMedian(0);
ca.setDevSum(mappedBases);
if(globalMean>-1) {ca.setDevSum((globalMean*globalMean)*length);}
if(windowAvg>0) {ca.setUnderWindowAverage(length);}
}
}
private void addToHist(CoverageArray ca, int strand, int length) {
LongList hist=(strand==0 ? hist0 : hist1);
assert(hist!=null) : stranded+", "+strand+", "+length;
if(ca==null) {
hist.increment(0, length);
return;
}
assert(ca.length()==length) : ca.length()+", "+length+", "+ca.maxIndex;
for(int i=0, len=ca.length(); i<len; i++) {
int d=Tools.min(ca.get(i), histMax);
hist.increment(d);
}
}
long totalSumT=0;
long binsT=0;
final int tid;
final int threads;
final ArrayList<Scaffold> list;
LongList hist0;
LongList hist1;
final boolean calcStd;
final boolean calcMedian;
final boolean calcHist;
final int minDepth;
final int histMax;
final double globalMean;
final double windowAvg;
final int window;
}
private static Scaffold getNext(ArrayBlockingQueue<Scaffold> queue) {
Scaffold next=null;
while(next==null) {
try {
next=queue.take();
} catch (InterruptedException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
return next;
}
private static void addToQueue(ArrayBlockingQueue<Scaffold> queue, Scaffold s) {
while(s!=null) {
try {
queue.put(s);
s=null;
} catch (InterruptedException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
}
@Override
public void accumulate(ProcessThread t) {
synchronized(t) {
totalSum+=t.totalSumT;
bins+=t.binsT;
if(hist0==null) {hist0=t.hist0;}
else if(t.hist0!=null) {hist0.incrementBy(t.hist0);}
if(hist1==null) {hist1=t.hist1;}
else if(t.hist1!=null) {hist1.incrementBy(t.hist1);}
}
}
@Override
public boolean success() {
// TODO Auto-generated method stub
return false;
}
/*--------------------------------------------------------------*/
/*---------------- Fields ----------------*/
/*--------------------------------------------------------------*/
long totalSum;
long bins;
double mean;
double stdev;
long totalSumBinned;
long binsBinned;
double meanBinned;
double stdevBinned;
public LongList hist0;
public LongList hist1;
@Override
public final ReadWriteLock rwlock() {return rwlock;}
private final ReadWriteLock rwlock=new ReentrantReadWriteLock();
public final int strand;
public final boolean stranded;
public boolean errorState=false;
private static final Scaffold POISON=new Scaffold("POISON", null, -1);
public static boolean verbose=false;
}
|