/**
* Checks whether CEL files and CDF files are from the same platform.
*
* @param normalCelFileName Affymetrix CEL file of the paired normal sample
* @param tumorCelFileName Affymetrix CEL file of the tumor sample
* @param cdfFileName Affymetrix library file (CDF file) for the platform on which the samples are
* generated
* @return <code>true</code> if the CEL files and CDF file are from the same platform, otherwise
* <code>false</code>
*/
private boolean checkChipType(
String normalCelFileName, String tumorCelFileName, String cdfFileName) {
// Check whether the normal CEL file, tumor CEL file, and the CDF file are consistent, and set
// the variable chipType.
FusionCELData celNormal;
FusionCELData celTumor;
FusionCDFData cdf;
celNormal = new FusionCELData();
celNormal.setFileName(normalCelFileName);
if (celNormal.read() == false) {
isSuccessfulOpenFile = false;
return false;
}
celTumor = new FusionCELData();
celTumor.setFileName(tumorCelFileName);
if (celTumor.read() == false) {
isSuccessfulOpenFile = false;
return false;
}
cdf = new FusionCDFData();
cdf.setFileName(cdfFileName);
if (cdf.read() == false) {
isSuccessfulOpenFile = false;
return false;
}
if (!celNormal.getChipType().equals(cdf.getChipType())) {
isSuccessfulOpenFile = false;
return false;
}
if (!celTumor.getChipType().equals(cdf.getChipType())) {
isSuccessfulOpenFile = false;
return false;
}
chipType = celNormal.getChipType();
isSuccessfulOpenFile = true;
return true;
}
/**
* Reads CEL files. <code>readCel</code> reads the intensity signals from the tumor sample and the
* normal sample. The copy number at a particular locus is calculated using the ratio of the tumor
* intensity at that locus and the correspoding normal intensity times 2.
*
* @param normalCelFileName Affymetrix CEL file of the paired normal sample
* @param tumorCelFileName Affymetrix CEL file of the tumor sample
* @param cdfFileName Affymetrix library file (CDF file) for the platform on which the samples are
* generated
*/
private void readCel(String normalCelFileName, String tumorCelFileName, String cdfFileName) {
FusionCELData celNormal;
FusionCELData celTumor;
FusionCDFData cdf;
celNormal = new FusionCELData();
celNormal.setFileName(normalCelFileName);
if (celNormal.read() == false) {
System.out.println("Failed to read the CEL file.");
return;
}
celTumor = new FusionCELData();
celTumor.setFileName(tumorCelFileName);
if (celTumor.read() == false) {
System.out.println("Failed to read the CEL file.");
return;
}
cdf = new FusionCDFData();
cdf.setFileName(cdfFileName);
if (cdf.read() == false) {
System.out.println("Failed to read the CDF file.");
return;
}
int nsets = cdf.getHeader().getNumProbeSets();
ProbeSetIntensityData[] probeSetDataNormal = new ProbeSetIntensityData[nsets];
ProbeSetIntensityData[] probeSetDataTumor = new ProbeSetIntensityData[nsets];
// SNP array structure:
// 1. Each probeset contains several groups (the number of groups varies among different
// probeset)
// 2. Each group contains several cells (the number of cells also varies among different
// groups)
for (int iset = 0; iset < nsets; iset++) {
String probeSetName = cdf.getProbeSetName(iset); // get the probeset name
FusionCDFProbeSetInformation set = new FusionCDFProbeSetInformation();
cdf.getProbeSetInformation(iset, set);
int ngroups = set.getNumGroups();
int numPmANormal = 0; // Pm : perfect match
int numPmBNormal = 0;
int numPmATumor = 0;
int numPmBTumor = 0;
int numMmANormal = 0; // Mm: Mis-Match
int numMmBNormal = 0;
int numMmATumor = 0;
int numMmBTumor = 0;
probeSetDataNormal[iset] = new ProbeSetIntensityData();
probeSetDataTumor[iset] = new ProbeSetIntensityData();
probeSetDataNormal[iset].probeSetType = set.getProbeSetType();
probeSetDataTumor[iset].probeSetType = set.getProbeSetType();
probeSetDataNormal[iset].probeSetID = cdf.getProbeSetName(iset) + "";
probeSetDataTumor[iset].probeSetID = cdf.getProbeSetName(iset) + "";
for (int igroup = 0; igroup < ngroups; igroup++) {
FusionCDFProbeGroupInformation group = new FusionCDFProbeGroupInformation();
set.getGroup(igroup, group);
int ncells = group.getNumCells();
for (int icell = 0; icell < ncells; icell++) {
FusionCDFProbeInformation probe = new FusionCDFProbeInformation();
group.getCell(icell, probe);
try {
char pBase = probe.getPBase();
char tBase = probe.getTBase();
// only If the match is perfect, the intensity of this cell
// contributes
if ((((pBase + tBase) == 213) || ((pBase + tBase) == 202))) {
// Perfect match is the match that with pBase:tBase = a:t or c:g
if ((igroup % 2) == 0) {
if (!celNormal.isOutlier(probe.getX(), probe.getY())) {
probeSetDataNormal[iset].pmA +=
celNormal.getIntensity(probe.getX(), probe.getY());
numPmANormal++;
}
if (!celTumor.isOutlier(probe.getX(), probe.getY())) {
probeSetDataTumor[iset].pmA += celTumor.getIntensity(probe.getX(), probe.getY());
numPmATumor++;
}
} else {
if (!celNormal.isOutlier(probe.getX(), probe.getY())) {
probeSetDataNormal[iset].pmB +=
celNormal.getIntensity(probe.getX(), probe.getY());
numPmBNormal++;
}
if (!celTumor.isOutlier(probe.getX(), probe.getY())) {
probeSetDataTumor[iset].pmB += celTumor.getIntensity(probe.getX(), probe.getY());
numPmBTumor++;
}
}
}
} catch (Exception e) {
}
}
}
// using the average of intensity of the perfect match cells as the intensity for certain
// probeset
if (numPmANormal != 0) {
probeSetDataNormal[iset].pmA = probeSetDataNormal[iset].pmA / numPmANormal;
}
if (numPmBNormal != 0) {
probeSetDataNormal[iset].pmB = probeSetDataNormal[iset].pmB / numPmBNormal;
}
if (numPmATumor != 0) {
probeSetDataTumor[iset].pmA = probeSetDataTumor[iset].pmA / numPmATumor;
}
if (numPmBTumor != 0) {
probeSetDataTumor[iset].pmB = probeSetDataTumor[iset].pmB / numPmBTumor;
}
}
try {
genotypeCalling(probeSetDataNormal); // to decide whether certain probe is AB type
} catch (Exception e) {
}
Arrays.sort(probeSetDataNormal);
Arrays.sort(probeSetDataTumor);
ArrayList<Integer> indexFoundList = new ArrayList<Integer>(probeSetDataNormal.length);
ArrayList<Integer> indexFoundListSNP = new ArrayList<Integer>(probeSetDataNormal.length);
for (int i = 0; i < numAnnotatedProbeSet; i++) {
int indexFound =
Arrays.binarySearch(probeSetDataNormal, new ProbeSetIntensityData(probeSetID[i]));
if (indexFound >= 0) {
isGenotypeAB[i] = probeSetDataNormal[indexFound].isGenotypeAB;
probeSetType[i] = probeSetDataNormal[indexFound].probeSetType;
copyNumber[i] =
2
* (probeSetDataTumor[indexFound].pmA + probeSetDataTumor[indexFound].pmB)
/ (probeSetDataNormal[indexFound].pmA + probeSetDataNormal[indexFound].pmB);
intensityNormal[i] =
probeSetDataNormal[indexFound].pmA + probeSetDataNormal[indexFound].pmB;
intensityTumor[i] = probeSetDataTumor[indexFound].pmA + probeSetDataTumor[indexFound].pmB;
// For 250k affymetrix chip, all the probes are SNP probes
if (probeSetType[i] == FusionGeneChipProbeSetType.GenotypingProbeSetType) {
alleleA[i] = probeSetDataTumor[indexFound].pmA / probeSetDataNormal[indexFound].pmA;
alleleB[i] = probeSetDataTumor[indexFound].pmB / probeSetDataNormal[indexFound].pmB;
if (alleleA[i] > 1E-10
&& alleleA[i] <= 30
&& alleleB[i] >= 1E-10
&& alleleB[i] <= 30
&& copyNumber[i] > 1E-10
&& copyNumber[i] <= 30) {
isOutlier[i] = false;
indexFoundList.add(i);
indexFoundListSNP.add(i);
}
}
// For SNP6.0, there are half SNP probes and half CN probes
if (probeSetType[i] == FusionGeneChipProbeSetType.CopyNumberProbeSetType) {
if (copyNumber[i] > 1E-10 && copyNumber[i] <= 30) {
isOutlier[i] = false;
indexFoundList.add(i);
}
}
}
}
/* Median filtering on the intensity data before calculating the copy numbers*/
// Filters f1 = new Filters(intensityNormal);
// f1.medianFilter(3);
// Filters f2 = new Filters(intensityTumor);
// f2.medianFilter(3);
// for (int j = 0; j < indexFoundList.size(); j ++) {
// int i = indexFoundList.get(j);
// copyNumber[i] = intensityTumor[i] / intensityNormal[i] * 2;
// if (!(copyNumber[i] > 1E-10 && copyNumber[i] <=30)) {
// copyNumber[i] = 2.0;
// isOutlier[i] = true;
// }
// }
globalNormalization(indexFoundList, indexFoundListSNP);
}