com.google.cloud.genomics.dataflow.functions.verifybamid.LikelihoodFn Maven / Gradle / Ivy
/*
* Copyright 2015 Google.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package com.google.cloud.genomics.dataflow.functions.verifybamid;
import com.google.cloud.genomics.dataflow.model.ReadCounts;
import com.google.cloud.genomics.dataflow.model.ReadQualityCount;
import com.google.cloud.genomics.dataflow.model.ReadQualityCount.Base;
import com.google.common.collect.ImmutableList;
import com.google.common.collect.ImmutableMap;
import com.google.genomics.v1.Position;
import org.apache.commons.math3.analysis.UnivariateFunction;
import java.util.ArrayList;
import java.util.Collections;
import java.util.Iterator;
import java.util.Map;
/**
* Implementation of the likelihood function in equation (2) in
* G. Jun, M. Flickinger, K. N. Hetrick, Kurt, J. M. Romm, K. F. Doheny,
* G. Abecasis, M. Boehnke,and H. M. Kang, Detecting and Estimating
* Contamination of Human DNA Samples in Sequencing and Array-Based Genotype
* Data, American journal of human genetics doi:10.1016/j.ajhg.2012.09.004
* (volume 91 issue 5 pp.839 - 848)
* http://www.sciencedirect.com/science/article/pii/S0002929712004788
*/
public class LikelihoodFn implements UnivariateFunction {
/** Possible genotypes for a SNP with a single alternate */
enum Genotype {
REF_HOMOZYGOUS, HETEROZYGOUS, NONREF_HOMOZYGOUS
}
/** Possible error statuses for a base in a read */
enum ReadStatus {
CORRECT, ERROR
}
static int toTableIndex(Base observed, Genotype trueGenotype, ReadStatus status) {
return observed.ordinal()
+ Base.values().length * (status.ordinal()
+ ReadStatus.values().length * trueGenotype.ordinal());
}
/*
* P_OBS_GIVEN_TRUTH contains the probability of observing a particular
* base (reference, non-reference, or other) given the true genotype
* and the error status of the read. See Table 1 of Jun et al.
*/
private static final ImmutableList P_OBS_GIVEN_TRUTH;
static {
final ImmutableList pTable = ImmutableList.of(
// Observed base
// REF NONREF OTHER
1.0, 0.0, 0.0, // P(base | REF_HOMOZYGOUS, CORRECT)
0.0, 1.0 / 3.0, 2.0 / 3.0, // P(base | REF_HOMOZYGOUS, ERROR)
0.5, 0.5, 0.0, // P(base | HETEROZYGOUS, CORRECT)
1.0 / 6.0, 1.0 / 6.0, 2.0 / 3.0, // P(base | HETEROZYGOUS, ERROR)
0.0, 1.0, 0.0, // P(base | NONREF_HOMOZYGOUS, CORRECT)
1.0 / 3.0, 0.0, 2.0 / 3.0); // P(base | NONREF_HOMOZYGOUS, ERROR)
Iterator itProb = pTable.iterator();
ArrayList pCond = new ArrayList<>();
pCond.addAll(Collections.nCopies(
Base.values().length * ReadStatus.values().length * Genotype.values().length, 0.0));
for (
Genotype g : ImmutableList.of(Genotype.REF_HOMOZYGOUS, Genotype.HETEROZYGOUS,
Genotype.NONREF_HOMOZYGOUS)) {
for (ReadStatus r : ImmutableList.of(ReadStatus.CORRECT, ReadStatus.ERROR)) {
for (Base b : ImmutableList.of(Base.REF, Base.NONREF, Base.OTHER)) {
pCond.set(toTableIndex(b, g, r), itProb.next());
}
}
}
P_OBS_GIVEN_TRUTH = ImmutableList.copyOf(pCond);
}
private final Map readCounts;
/**
* Create a new LikelihoodFn instance for a given set of read counts.
*
* @param readCounts counts of reads by quality for each position of interest
*/
public LikelihoodFn(Map readCounts) {
// copy the map so the counts don't get changed out from under us
this.readCounts = ImmutableMap.copyOf(readCounts);
}
/**
* Compute the probability of a genotype given the reference allele probability.
*/
private static double pGenotype(Genotype g, double refProb) {
switch(g) {
case REF_HOMOZYGOUS:
return refProb * refProb;
case HETEROZYGOUS:
return refProb * (1.0 - refProb);
case NONREF_HOMOZYGOUS:
return (1.0 - refProb) * (1.0 - refProb);
default:
throw new IllegalArgumentException("Illegal genotype");
}
}
/**
* Look up the probability of an observation conditioned on the underlying state.
*/
private static double probObsGivenTruth(Base observed, Genotype trueGenotype,
ReadStatus trueStatus) {
return P_OBS_GIVEN_TRUTH.get(toTableIndex(observed, trueGenotype, trueStatus));
}
/**
* Compute the likelihood of a contaminant fraction alpha.
*
* See equation (2) in Jun et al.
*/
@Override
public double value(double alpha) {
double logLikelihood = 0.0;
for (ReadCounts rc : readCounts.values()) {
double refProb = rc.getRefFreq();
double pPosition = 0.0;
for (Genotype trueGenotype1 : Genotype.values()) {
double pGenotype1 = pGenotype(trueGenotype1, refProb);
for (Genotype trueGenotype2 : Genotype.values()) {
double pGenotype2 = pGenotype(trueGenotype2, refProb);
double pObsGivenGenotype = 1.0;
for (ReadQualityCount rqc : rc.getReadQualityCounts()) {
Base base = rqc.getBase();
double pErr = phredToProb(rqc.getQuality());
double pObs
= ((1.0 - alpha)
* probObsGivenTruth(base, trueGenotype1, ReadStatus.CORRECT)
+ (alpha)
* probObsGivenTruth(base, trueGenotype2, ReadStatus.CORRECT)
) * (1.0 - pErr)
+ ((1.0 - alpha)
* probObsGivenTruth(base, trueGenotype1, ReadStatus.ERROR)
+ (alpha)
* probObsGivenTruth(base, trueGenotype2, ReadStatus.ERROR)
) * pErr;
pObsGivenGenotype *= Math.pow(pObs, rqc.getCount());
}
pPosition += pObsGivenGenotype * pGenotype1 * pGenotype2;
}
}
logLikelihood += Math.log(pPosition);
}
return logLikelihood;
}
/**
* Convert a Phred score to a probability.
*/
private static double phredToProb(int phred) {
return Math.pow(10.0, -(double) phred / 10.0);
}
}