707 lines
37 KiB
Java
707 lines
37 KiB
Java
import org.jgrapht.alg.interfaces.MatchingAlgorithm;
|
|
import org.jgrapht.alg.matching.MaximumWeightBipartiteMatching;
|
|
import org.jgrapht.generate.SimpleWeightedBipartiteGraphMatrixGenerator;
|
|
import org.jgrapht.graph.DefaultWeightedEdge;
|
|
import org.jgrapht.graph.SimpleWeightedGraph;
|
|
import org.jheaps.tree.PairingHeap;
|
|
|
|
import java.math.BigDecimal;
|
|
import java.math.MathContext;
|
|
import java.text.NumberFormat;
|
|
import java.time.Instant;
|
|
import java.time.Duration;
|
|
import java.util.*;
|
|
import java.util.stream.IntStream;
|
|
|
|
//NOTE: "sequence" in method and variable names refers to a peptide sequence from a simulated T cell
|
|
public class Simulator {
|
|
private static final int cdr3AlphaIndex = 0;
|
|
private static final int cdr3BetaIndex = 1;
|
|
private static final int cdr1AlphaIndex = 2;
|
|
private static final int cdr1BetaIndex = 3;
|
|
|
|
public static CellSample generateCellSample(Integer numDistinctCells, Integer cdr1Freq) {
|
|
//In real T cells, CDR1s have about one third the diversity of CDR3s
|
|
List<Integer> numbersCDR3 = new ArrayList<>();
|
|
List<Integer> numbersCDR1 = new ArrayList<>();
|
|
Integer numDistCDR3s = 2 * numDistinctCells + 1;
|
|
IntStream.range(1, numDistCDR3s + 1).forEach(i -> numbersCDR3.add(i));
|
|
IntStream.range(numDistCDR3s + 1, numDistCDR3s + 1 + (numDistCDR3s / cdr1Freq) + 1).forEach(i -> numbersCDR1.add(i));
|
|
Collections.shuffle(numbersCDR3);
|
|
Collections.shuffle(numbersCDR1);
|
|
|
|
//Each cell represented by 4 values
|
|
//two CDR3s, and two CDR1s. First two values are CDR3s (alpha, beta), second two are CDR1s (alpha, beta)
|
|
List<Integer[]> distinctCells = new ArrayList<>();
|
|
for(int i = 0; i < numbersCDR3.size() - 1; i = i + 2){
|
|
Integer tmpCDR3a = numbersCDR3.get(i);
|
|
Integer tmpCDR3b = numbersCDR3.get(i+1);
|
|
Integer tmpCDR1a = numbersCDR1.get(i % numbersCDR1.size());
|
|
Integer tmpCDR1b = numbersCDR1.get((i+1) % numbersCDR1.size());
|
|
Integer[] tmp = {tmpCDR3a, tmpCDR3b, tmpCDR1a, tmpCDR1b};
|
|
distinctCells.add(tmp);
|
|
}
|
|
return new CellSample(distinctCells, cdr1Freq);
|
|
}
|
|
|
|
//Make the graph needed for matching CDR3s
|
|
public static GraphWithMapData makeGraph(List<Integer[]> distinctCells, Plate samplePlate, Integer lowThreshold,
|
|
Integer highThreshold, boolean verbose) {
|
|
Instant start = Instant.now();
|
|
int[] alphaIndex = {cdr3AlphaIndex};
|
|
int[] betaIndex = {cdr3BetaIndex};
|
|
int numWells = samplePlate.getSize();
|
|
|
|
if(verbose){System.out.println("Making cell maps");}
|
|
//HashMap keyed to Alphas, values Betas
|
|
Map<Integer, Integer> distCellsMapAlphaKey = makeSequenceToSequenceMap(distinctCells, 0, 1);
|
|
if(verbose){System.out.println("Cell maps made");}
|
|
|
|
if(verbose){System.out.println("Making well maps");}
|
|
Map<Integer, Integer> allAlphas = samplePlate.assayWellsPeptideP(alphaIndex);
|
|
Map<Integer, Integer> allBetas = samplePlate.assayWellsPeptideP(betaIndex);
|
|
int alphaCount = allAlphas.size();
|
|
if(verbose){System.out.println("All alphas count: " + alphaCount);}
|
|
int betaCount = allBetas.size();
|
|
if(verbose){System.out.println("All betas count: " + betaCount);}
|
|
|
|
if(verbose){System.out.println("Well maps made");}
|
|
|
|
//Remove saturating-occupancy sequences because they have no signal value.
|
|
//Remove sequences with total occupancy below minimum pair overlap threshold
|
|
if(verbose){System.out.println("Removing sequences present in all wells.");}
|
|
if(verbose){System.out.println("Removing sequences with occupancy below the minimum overlap threshold");}
|
|
filterByOccupancyThreshold(allAlphas, lowThreshold, numWells - 1);
|
|
filterByOccupancyThreshold(allBetas, lowThreshold, numWells - 1);
|
|
if(verbose){System.out.println("Sequences removed");}
|
|
int pairableAlphaCount = allAlphas.size();
|
|
if(verbose){System.out.println("Remaining alphas count: " + pairableAlphaCount);}
|
|
int pairableBetaCount = allBetas.size();
|
|
if(verbose){System.out.println("Remaining betas count: " + pairableBetaCount);}
|
|
|
|
if(verbose){System.out.println("Making vertex maps");}
|
|
//For the SimpleWeightedBipartiteGraphMatrixGenerator, all vertices must have
|
|
//distinct numbers associated with them. Since I'm using a 2D array, that means
|
|
//distinct indices between the rows and columns. vertexStartValue lets me track where I switch
|
|
//from numbering rows to columns, so I can assign unique numbers to every vertex, and then
|
|
//subtract the vertexStartValue from betas to use their vertex labels as array indices
|
|
Integer vertexStartValue = 0;
|
|
//keys are sequential integer vertices, values are alphas
|
|
Map<Integer, Integer> plateVtoAMap = makeVertexToSequenceMap(allAlphas, vertexStartValue);
|
|
//new start value for vertex to beta map should be one more than final vertex value in alpha map
|
|
vertexStartValue += plateVtoAMap.size();
|
|
//keys are sequential integers vertices, values are betas
|
|
Map<Integer, Integer> plateVtoBMap = makeVertexToSequenceMap(allBetas, vertexStartValue);
|
|
//keys are alphas, values are sequential integer vertices from previous map
|
|
Map<Integer, Integer> plateAtoVMap = invertVertexMap(plateVtoAMap);
|
|
//keys are betas, values are sequential integer vertices from previous map
|
|
Map<Integer, Integer> plateBtoVMap = invertVertexMap(plateVtoBMap);
|
|
if(verbose){System.out.println("Vertex maps made");}
|
|
|
|
//make adjacency matrix for bipartite graph generator
|
|
//(technically this is only 1/4 of an adjacency matrix, but that's all you need
|
|
//for a bipartite graph, and all the SimpleWeightedBipartiteGraphMatrixGenerator class expects.)
|
|
if(verbose){System.out.println("Creating adjacency matrix");}
|
|
//Count how many wells each alpha appears in
|
|
Map<Integer, Integer> alphaWellCounts = new HashMap<>();
|
|
//count how many wells each beta appears in
|
|
Map<Integer, Integer> betaWellCounts = new HashMap<>();
|
|
//the adjacency matrix to be used by the graph generator
|
|
double[][] weights = new double[plateVtoAMap.size()][plateVtoBMap.size()];
|
|
countSequencesAndFillMatrix(samplePlate, allAlphas, allBetas, plateAtoVMap,
|
|
plateBtoVMap, alphaIndex, betaIndex, alphaWellCounts, betaWellCounts, weights);
|
|
if(verbose){System.out.println("Matrix created");}
|
|
|
|
//create bipartite graph
|
|
if(verbose){System.out.println("Creating graph");}
|
|
//the graph object
|
|
SimpleWeightedGraph<Integer, DefaultWeightedEdge> graph =
|
|
new SimpleWeightedGraph<>(DefaultWeightedEdge.class);
|
|
//the graph generator
|
|
SimpleWeightedBipartiteGraphMatrixGenerator graphGenerator = new SimpleWeightedBipartiteGraphMatrixGenerator();
|
|
//the list of alpha vertices
|
|
List<Integer> alphaVertices = new ArrayList<>(plateVtoAMap.keySet()); //This will work because LinkedHashMap preserves order of entry
|
|
graphGenerator.first(alphaVertices);
|
|
//the list of beta vertices
|
|
List<Integer> betaVertices = new ArrayList<>(plateVtoBMap.keySet());
|
|
graphGenerator.second(betaVertices); //This will work because LinkedHashMap preserves order of entry
|
|
//use adjacency matrix of weight created previously
|
|
graphGenerator.weights(weights);
|
|
graphGenerator.generateGraph(graph);
|
|
if(verbose){System.out.println("Graph created");}
|
|
|
|
//remove weights outside given overlap thresholds
|
|
if(verbose){System.out.println("Eliminating edges with weights outside overlap threshold values");}
|
|
filterByOccupancyThreshold(graph, lowThreshold, highThreshold);
|
|
if(verbose){System.out.println("Over- and under-weight edges set to 0.0");}
|
|
Instant stop = Instant.now();
|
|
Duration time = Duration.between(start, stop);
|
|
//return GraphWithMapData object
|
|
return new GraphWithMapData(graph, numWells, samplePlate.getConcentrations(), alphaCount, betaCount, lowThreshold, highThreshold,
|
|
distCellsMapAlphaKey, plateVtoAMap, plateVtoBMap, plateAtoVMap,
|
|
plateBtoVMap, alphaWellCounts, betaWellCounts, time);
|
|
}
|
|
|
|
//match CDR3s.
|
|
public static MatchingResult matchCDR3s(GraphWithMapData data, Integer maxOccupancyDifference,
|
|
Integer minOverlapPercent, boolean verbose) {
|
|
Instant start = Instant.now();
|
|
int numWells = data.getNumWells();
|
|
Integer highThreshold = data.getHighThreshold();
|
|
Integer lowThreshold = data.getLowThreshold();
|
|
Integer alphaCount = data.getAlphaCount();
|
|
Integer betaCount = data.getBetaCount();
|
|
Map<Integer, Integer> distCellsMapAlphaKey = data.getDistCellsMapAlphaKey();
|
|
Map<Integer, Integer> plateVtoAMap = data.getPlateVtoAMap();
|
|
Map<Integer, Integer> plateVtoBMap = data.getPlateVtoBMap();
|
|
Map<Integer, Integer> alphaWellCounts = data.getAlphaWellCounts();
|
|
Map<Integer, Integer> betaWellCounts = data.getBetaWellCounts();
|
|
SimpleWeightedGraph<Integer, DefaultWeightedEdge> graph = data.getGraph();
|
|
|
|
//Filter by overlap size
|
|
if(verbose){System.out.println("Eliminating edges with weights less than " + minOverlapPercent.toString() +
|
|
" percent of vertex occupancy value.");}
|
|
filterByOverlapSize(graph, alphaWellCounts, betaWellCounts, plateVtoAMap, plateVtoBMap, minOverlapPercent);
|
|
if(verbose){System.out.println("Edges with weights too far below vertex occupancy values set to 0.0");}
|
|
|
|
//Filter by relative occupancy
|
|
if(verbose){System.out.println("Eliminating edges between vertices with occupancy difference > "
|
|
+ maxOccupancyDifference);}
|
|
filterByRelativeOccupancy(graph, alphaWellCounts, betaWellCounts, plateVtoAMap, plateVtoBMap,
|
|
maxOccupancyDifference);
|
|
if(verbose){System.out.println("Edges between vertices of with excessively different occupancy values " +
|
|
"set to 0.0");}
|
|
//Find Maximum Weighted Matching
|
|
//using jheaps library class PairingHeap for improved efficiency
|
|
if(verbose){System.out.println("Finding maximum weighted matching");}
|
|
//Attempting to use addressable heap to improve performance
|
|
MaximumWeightBipartiteMatching maxWeightMatching =
|
|
new MaximumWeightBipartiteMatching(graph,
|
|
plateVtoAMap.keySet(),
|
|
plateVtoBMap.keySet(),
|
|
i -> new PairingHeap(Comparator.naturalOrder()));
|
|
MatchingAlgorithm.Matching<String, DefaultWeightedEdge> graphMatching = maxWeightMatching.getMatching();
|
|
if(verbose){System.out.println("Matching completed");}
|
|
Instant stop = Instant.now();
|
|
|
|
//Header for CSV file
|
|
List<String> header = new ArrayList<>();
|
|
header.add("Alpha");
|
|
header.add("Alpha well count");
|
|
header.add("Beta");
|
|
header.add("Beta well count");
|
|
header.add("Overlap well count");
|
|
header.add("Matched correctly?");
|
|
header.add("P-value");
|
|
|
|
//Results for csv file
|
|
List<List<String>> allResults = new ArrayList<>();
|
|
NumberFormat nf = NumberFormat.getInstance(Locale.US);
|
|
MathContext mc = new MathContext(3);
|
|
Iterator<DefaultWeightedEdge> weightIter = graphMatching.iterator();
|
|
DefaultWeightedEdge e;
|
|
int trueCount = 0;
|
|
int falseCount = 0;
|
|
boolean check;
|
|
Map<Integer, Integer> matchMap = new HashMap<>();
|
|
while(weightIter.hasNext()) {
|
|
e = weightIter.next();
|
|
Integer source = graph.getEdgeSource(e);
|
|
Integer target = graph.getEdgeTarget(e);
|
|
//The match map is all matches found, not just true matches!
|
|
matchMap.put(plateVtoAMap.get(source), plateVtoBMap.get(target));
|
|
check = plateVtoBMap.get(target).equals(distCellsMapAlphaKey.get(plateVtoAMap.get(source)));
|
|
if(check) {
|
|
trueCount++;
|
|
}
|
|
else {
|
|
falseCount++;
|
|
}
|
|
List<String> result = new ArrayList<>();
|
|
result.add(plateVtoAMap.get(source).toString());
|
|
//alpha well count
|
|
result.add(alphaWellCounts.get(plateVtoAMap.get(source)).toString());
|
|
result.add(plateVtoBMap.get(target).toString());
|
|
//beta well count
|
|
result.add(betaWellCounts.get(plateVtoBMap.get(target)).toString());
|
|
//overlap count
|
|
result.add(Double.toString(graph.getEdgeWeight(e)));
|
|
result.add(Boolean.toString(check));
|
|
double pValue = Equations.pValue(numWells, alphaWellCounts.get(plateVtoAMap.get(source)),
|
|
betaWellCounts.get(plateVtoBMap.get(target)), graph.getEdgeWeight(e));
|
|
BigDecimal pValueTrunc = new BigDecimal(pValue, mc);
|
|
result.add(pValueTrunc.toString());
|
|
allResults.add(result);
|
|
}
|
|
|
|
//Metadate comments for CSV file
|
|
int min = Math.min(alphaCount, betaCount);
|
|
double attemptRate = (double) (trueCount + falseCount) / min;
|
|
BigDecimal attemptRateTrunc = new BigDecimal(attemptRate, mc);
|
|
double pairingErrorRate = (double) falseCount / (trueCount + falseCount);
|
|
BigDecimal pairingErrorRateTrunc = new BigDecimal(pairingErrorRate, mc);
|
|
//make list of well concentrations
|
|
List<Integer> wellConcentrations = Arrays.asList(data.getWellConcentrations());
|
|
//make string out of concentrations list
|
|
StringBuilder concen = new StringBuilder();
|
|
for(Integer i: wellConcentrations){
|
|
concen.append(i.toString());
|
|
concen.append(" ");
|
|
}
|
|
String concenString = concen.toString();
|
|
|
|
List<String> comments = new ArrayList<>();
|
|
comments.add("T cell counts in sample plate wells: " + concenString);
|
|
comments.add("Total alphas found: " + alphaCount);
|
|
comments.add("Total betas found: " + betaCount);
|
|
comments.add("High overlap threshold: " + highThreshold);
|
|
comments.add("Low overlap threshold: " + lowThreshold);
|
|
comments.add("Minimum overlap percent: " + minOverlapPercent);
|
|
comments.add("Maximum occupancy difference: " + maxOccupancyDifference);
|
|
comments.add("Pairing attempt rate: " + attemptRateTrunc);
|
|
comments.add("Correct pairings: " + trueCount);
|
|
comments.add("Incorrect pairings: " + falseCount);
|
|
comments.add("Pairing error rate: " + pairingErrorRateTrunc);
|
|
Duration time = Duration.between(start, stop);
|
|
time = time.plus(data.getTime());
|
|
comments.add("Simulation time: " + nf.format(time.toSeconds()) + " seconds");
|
|
if(verbose){
|
|
for(String s: comments){
|
|
System.out.println(s);
|
|
}
|
|
}
|
|
return new MatchingResult(data.getSourceFilename(), comments, header, allResults, matchMap, time);
|
|
}
|
|
|
|
//Simulated matching of CDR1s to CDR3s. Requires MatchingResult from prior run of matchCDR3s.
|
|
public static MatchingResult[] matchCDR1s(List<Integer[]> distinctCells,
|
|
Plate samplePlate, Integer lowThreshold,
|
|
Integer highThreshold, MatchingResult priorResult){
|
|
Instant start = Instant.now();
|
|
Duration previousTime = priorResult.getTime();
|
|
Map<Integer, Integer> previousMatches = priorResult.getMatchMap();
|
|
int numWells = samplePlate.getSize();
|
|
int[] cdr3Indices = {cdr3AlphaIndex, cdr3BetaIndex};
|
|
int[] cdr1Indices = {cdr1AlphaIndex, cdr1BetaIndex};
|
|
|
|
System.out.println("Making previous match maps");
|
|
Map<Integer, Integer> cdr3AtoBMap = previousMatches;
|
|
Map<Integer, Integer> cdr3BtoAMap = invertVertexMap(cdr3AtoBMap);
|
|
System.out.println("Previous match maps made");
|
|
|
|
System.out.println("Making cell maps");
|
|
Map<Integer, Integer> alphaCDR3toCDR1Map = makeSequenceToSequenceMap(distinctCells, cdr3AlphaIndex, cdr1AlphaIndex);
|
|
Map<Integer, Integer> betaCDR3toCDR1Map = makeSequenceToSequenceMap(distinctCells, cdr3BetaIndex, cdr1BetaIndex);
|
|
System.out.println("Cell maps made");
|
|
|
|
System.out.println("Making well maps");
|
|
Map<Integer, Integer> allCDR3s = samplePlate.assayWellsPeptideP(cdr3Indices);
|
|
Map<Integer, Integer> allCDR1s = samplePlate.assayWellsPeptideP(cdr1Indices);
|
|
int CDR3Count = allCDR3s.size();
|
|
System.out.println("all CDR3s count: " + CDR3Count);
|
|
int CDR1Count = allCDR1s.size();
|
|
System.out.println("all CDR1s count: " + CDR1Count);
|
|
System.out.println("Well maps made");
|
|
|
|
System.out.println("Removing unpaired CDR3s from well maps");
|
|
List<Integer> unpairedCDR3s = new ArrayList<>();
|
|
for(Integer i: allCDR3s.keySet()){
|
|
if(!(cdr3AtoBMap.containsKey(i) || cdr3BtoAMap.containsKey(i))){
|
|
unpairedCDR3s.add(i);
|
|
}
|
|
}
|
|
for(Integer i: unpairedCDR3s){
|
|
allCDR3s.remove(i);
|
|
}
|
|
System.out.println("Unpaired CDR3s removed.");
|
|
System.out.println("Remaining CDR3 count: " + allCDR3s.size());
|
|
|
|
System.out.println("Removing below-minimum-overlap-threshold and saturating-occupancy CDR1s");
|
|
filterByOccupancyThreshold(allCDR1s, lowThreshold, numWells - 1);
|
|
System.out.println("CDR1s removed.");
|
|
System.out.println("Remaining CDR1 count: " + allCDR1s.size());
|
|
|
|
System.out.println("Making vertex maps");
|
|
|
|
//For the SimpleWeightedBipartiteGraphMatrixGenerator, all vertices must have
|
|
// distinct numbers associated with them. Since I'm using a 2D array, that means
|
|
// distinct indices between the rows and columns. vertexStartValue lets me track where I switch
|
|
// from numbering rows to columns, so I can assign unique numbers to every vertex, and then
|
|
// subtract the vertexStartValue from CDR1s to use their vertex labels as array indices
|
|
Integer vertexStartValue = 0;
|
|
//keys are sequential integer vertices, values are CDR3s
|
|
Map<Integer, Integer> plateVtoCDR3Map = makeVertexToSequenceMap(allCDR3s, vertexStartValue);
|
|
//New start value for vertex to CDR1 map should be one more than final vertex value in CDR3 map
|
|
vertexStartValue += plateVtoCDR3Map.size();
|
|
//keys are sequential integers vertices, values are CDR1s
|
|
Map<Integer, Integer> plateVtoCDR1Map = makeVertexToSequenceMap(allCDR1s, vertexStartValue);
|
|
//keys are CDR3s, values are sequential integer vertices from previous map
|
|
Map<Integer, Integer> plateCDR3toVMap = invertVertexMap(plateVtoCDR3Map);
|
|
//keys are CDR1s, values are sequential integer vertices from previous map
|
|
Map<Integer, Integer> plateCDR1toVMap = invertVertexMap(plateVtoCDR1Map);
|
|
System.out.println("Vertex maps made");
|
|
|
|
System.out.println("Creating adjacency matrix");
|
|
//Count how many wells each CDR3 appears in
|
|
Map<Integer, Integer> cdr3WellCounts = new HashMap<>();
|
|
//count how many wells each CDR1 appears in
|
|
Map<Integer, Integer> cdr1WellCounts = new HashMap<>();
|
|
//add edges, where weights are number of wells the peptides share in common.
|
|
//If this is too slow, can make a 2d array and use the SimpleWeightedGraphMatrixGenerator class
|
|
Map<Integer, Integer> wellNCDR3s = null;
|
|
Map<Integer, Integer> wellNCDR1s = null;
|
|
double[][] weights = new double[plateVtoCDR3Map.size()][plateVtoCDR1Map.size()];
|
|
countSequencesAndFillMatrix(samplePlate, allCDR3s, allCDR1s, plateCDR3toVMap, plateCDR1toVMap,
|
|
cdr3Indices, cdr1Indices, cdr3WellCounts, cdr1WellCounts, weights);
|
|
System.out.println("Matrix created");
|
|
|
|
System.out.println("Creating graph");
|
|
SimpleWeightedGraph<Integer, DefaultWeightedEdge> graph =
|
|
new SimpleWeightedGraph<>(DefaultWeightedEdge.class);
|
|
|
|
SimpleWeightedBipartiteGraphMatrixGenerator graphGenerator = new SimpleWeightedBipartiteGraphMatrixGenerator();
|
|
List<Integer> cdr3Vertices = new ArrayList<>(plateVtoCDR3Map.keySet()); //This will work because LinkedHashMap preserves order of entry
|
|
graphGenerator.first(cdr3Vertices);
|
|
List<Integer> cdr1Vertices = new ArrayList<>(plateVtoCDR1Map.keySet());
|
|
graphGenerator.second(cdr1Vertices); //This will work because LinkedHashMap preserves order of entry
|
|
graphGenerator.weights(weights);
|
|
graphGenerator.generateGraph(graph);
|
|
System.out.println("Graph created");
|
|
|
|
System.out.println("Removing edges outside of weight thresholds");
|
|
filterByOccupancyThreshold(graph, lowThreshold, highThreshold);
|
|
System.out.println("Over- and under-weight edges set to 0.0");
|
|
|
|
System.out.println("Finding first maximum weighted matching");
|
|
MaximumWeightBipartiteMatching firstMaxWeightMatching =
|
|
new MaximumWeightBipartiteMatching(graph, plateVtoCDR3Map.keySet(), plateVtoCDR1Map.keySet());
|
|
MatchingAlgorithm.Matching<String, DefaultWeightedEdge> graphMatching = firstMaxWeightMatching.getMatching();
|
|
System.out.println("First maximum weighted matching found");
|
|
|
|
|
|
//first processing run
|
|
Map<Integer, Integer> firstMatchCDR3toCDR1Map = new HashMap<>();
|
|
Iterator<DefaultWeightedEdge> weightIter = graphMatching.iterator();
|
|
DefaultWeightedEdge e;
|
|
while(weightIter.hasNext()){
|
|
e = weightIter.next();
|
|
// if(graph.getEdgeWeight(e) < lowThreshold || graph.getEdgeWeight(e) > highThreshold) {
|
|
// continue;
|
|
// }
|
|
Integer source = graph.getEdgeSource(e);
|
|
Integer target = graph.getEdgeTarget(e);
|
|
firstMatchCDR3toCDR1Map.put(plateVtoCDR3Map.get(source), plateVtoCDR1Map.get(target));
|
|
}
|
|
System.out.println("First pass matches: " + firstMatchCDR3toCDR1Map.size());
|
|
|
|
System.out.println("Removing edges from first maximum weighted matching");
|
|
//zero out the edge weights in the matching
|
|
weightIter = graphMatching.iterator();
|
|
while(weightIter.hasNext()){
|
|
graph.removeEdge(weightIter.next());
|
|
}
|
|
System.out.println("Edges removed");
|
|
|
|
//Generate a new matching
|
|
System.out.println("Finding second maximum weighted matching");
|
|
MaximumWeightBipartiteMatching secondMaxWeightMatching =
|
|
new MaximumWeightBipartiteMatching(graph, plateVtoCDR3Map.keySet(), plateVtoCDR1Map.keySet());
|
|
graphMatching = secondMaxWeightMatching.getMatching();
|
|
System.out.println("Second maximum weighted matching found");
|
|
|
|
|
|
//second processing run
|
|
Map<Integer, Integer> secondMatchCDR3toCDR1Map = new HashMap<>();
|
|
weightIter = graphMatching.iterator();
|
|
while(weightIter.hasNext()){
|
|
e = weightIter.next();
|
|
// if(graph.getEdgeWeight(e) < lowThreshold || graph.getEdgeWeight(e) > highThreshold) {
|
|
// continue;
|
|
// }
|
|
Integer source = graph.getEdgeSource(e);
|
|
// if(!(CDR3AtoBMap.containsKey(source) || CDR3BtoAMap.containsKey(source))){
|
|
// continue;
|
|
// }
|
|
Integer target = graph.getEdgeTarget(e);
|
|
secondMatchCDR3toCDR1Map.put(plateVtoCDR3Map.get(source), plateVtoCDR1Map.get(target));
|
|
}
|
|
System.out.println("Second pass matches: " + secondMatchCDR3toCDR1Map.size());
|
|
|
|
System.out.println("Mapping first pass CDR3 alpha/beta pairs");
|
|
//get linked map for first matching attempt
|
|
Map<Integer, Integer> firstMatchesMap = new LinkedHashMap<>();
|
|
for(Integer alphaCDR3: cdr3AtoBMap.keySet()) {
|
|
if (!(firstMatchCDR3toCDR1Map.containsKey(alphaCDR3))) {
|
|
continue;
|
|
}
|
|
Integer betaCDR3 = cdr3AtoBMap.get(alphaCDR3);
|
|
if (!(firstMatchCDR3toCDR1Map.containsKey(betaCDR3))) {
|
|
continue;
|
|
}
|
|
firstMatchesMap.put(alphaCDR3, firstMatchCDR3toCDR1Map.get(alphaCDR3));
|
|
firstMatchesMap.put(betaCDR3, firstMatchCDR3toCDR1Map.get(betaCDR3));
|
|
}
|
|
System.out.println("First pass CDR3 alpha/beta pairs mapped");
|
|
|
|
System.out.println("Mapping second pass CDR3 alpha/beta pairs.");
|
|
System.out.println("Finding CDR3 pairs that swapped CDR1 matches between first pass and second pass.");
|
|
//Look for matches that simply swapped already-matched alpha and beta CDR3s
|
|
Map<Integer, Integer> dualMatchesMap = new LinkedHashMap<>();
|
|
for(Integer alphaCDR3: cdr3AtoBMap.keySet()) {
|
|
if (!(firstMatchCDR3toCDR1Map.containsKey(alphaCDR3) && secondMatchCDR3toCDR1Map.containsKey(alphaCDR3))) {
|
|
continue;
|
|
}
|
|
Integer betaCDR3 = cdr3AtoBMap.get(alphaCDR3);
|
|
if (!(firstMatchCDR3toCDR1Map.containsKey(betaCDR3) && secondMatchCDR3toCDR1Map.containsKey(betaCDR3))) {
|
|
continue;
|
|
}
|
|
if(firstMatchCDR3toCDR1Map.get(alphaCDR3).equals(secondMatchCDR3toCDR1Map.get(betaCDR3))){
|
|
if(firstMatchCDR3toCDR1Map.get(betaCDR3).equals(secondMatchCDR3toCDR1Map.get(alphaCDR3))){
|
|
dualMatchesMap.put(alphaCDR3, firstMatchCDR3toCDR1Map.get(alphaCDR3));
|
|
dualMatchesMap.put(betaCDR3, firstMatchCDR3toCDR1Map.get(betaCDR3));
|
|
}
|
|
}
|
|
}
|
|
System.out.println("Second pass mapping made. Dual CDR3/CDR1 pairings found.");
|
|
|
|
Instant stop = Instant.now();
|
|
//results for first map
|
|
System.out.println("RESULTS FOR FIRST PASS MATCHING");
|
|
List<List<String>> allResults = new ArrayList<>();
|
|
Integer trueCount = 0;
|
|
Iterator iter = firstMatchesMap.keySet().iterator();
|
|
|
|
while(iter.hasNext()){
|
|
Boolean proven = false;
|
|
List<String> tmp = new ArrayList<>();
|
|
tmp.add(iter.next().toString());
|
|
tmp.add(iter.next().toString());
|
|
tmp.add(firstMatchesMap.get(Integer.valueOf(tmp.get(0))).toString());
|
|
tmp.add(firstMatchesMap.get(Integer.valueOf(tmp.get(1))).toString());
|
|
if(alphaCDR3toCDR1Map.get(Integer.valueOf(tmp.get(0))).equals(Integer.valueOf(tmp.get(2)))){
|
|
if(betaCDR3toCDR1Map.get(Integer.valueOf(tmp.get(1))).equals(Integer.valueOf(tmp.get(3)))){
|
|
proven = true;
|
|
}
|
|
}
|
|
else if(alphaCDR3toCDR1Map.get(Integer.valueOf(tmp.get(0))).equals(Integer.valueOf(tmp.get(3)))){
|
|
if(betaCDR3toCDR1Map.get(Integer.valueOf(tmp.get(1))).equals(Integer.valueOf(tmp.get(2)))){
|
|
proven = true;
|
|
}
|
|
}
|
|
tmp.add(proven.toString());
|
|
allResults.add(tmp);
|
|
if(proven){
|
|
trueCount++;
|
|
}
|
|
}
|
|
|
|
List<String> comments = new ArrayList<>();
|
|
comments.add("Plate size: " + samplePlate.getSize() + " wells");
|
|
comments.add("Previous pairs found: " + previousMatches.size());
|
|
comments.add("CDR1 matches attempted: " + allResults.size());
|
|
double attemptRate = (double) allResults.size() / previousMatches.size();
|
|
comments.add("Matching attempt rate: " + attemptRate);
|
|
comments.add("Number of correct matches: " + trueCount);
|
|
double correctRate = (double) trueCount / allResults.size();
|
|
comments.add("Correct matching rate: " + correctRate);
|
|
NumberFormat nf = NumberFormat.getInstance(Locale.US);
|
|
Duration time = Duration.between(start, stop);
|
|
time = time.plus(previousTime);
|
|
comments.add("Simulation time: " + nf.format(time.toSeconds()) + " seconds");
|
|
for(String s: comments){
|
|
System.out.println(s);
|
|
}
|
|
|
|
|
|
|
|
List<String> headers = new ArrayList<>();
|
|
headers.add("CDR3 alpha");
|
|
headers.add("CDR3 beta");
|
|
headers.add("first matched CDR1");
|
|
headers.add("second matched CDR1");
|
|
headers.add("Correct match?");
|
|
|
|
MatchingResult firstTest = new MatchingResult(samplePlate.getSourceFileName(),
|
|
comments, headers, allResults, dualMatchesMap, time);
|
|
|
|
//results for dual map
|
|
System.out.println("RESULTS FOR SECOND PASS MATCHING");
|
|
allResults = new ArrayList<>();
|
|
trueCount = 0;
|
|
iter = dualMatchesMap.keySet().iterator();
|
|
while(iter.hasNext()){
|
|
Boolean proven = false;
|
|
List<String> tmp = new ArrayList<>();
|
|
tmp.add(iter.next().toString());
|
|
tmp.add(iter.next().toString());
|
|
tmp.add(dualMatchesMap.get(Integer.valueOf(tmp.get(0))).toString());
|
|
tmp.add(dualMatchesMap.get(Integer.valueOf(tmp.get(1))).toString());
|
|
if(alphaCDR3toCDR1Map.get(Integer.valueOf(tmp.get(0))).equals(Integer.valueOf(tmp.get(2)))){
|
|
if(betaCDR3toCDR1Map.get(Integer.valueOf(tmp.get(1))).equals(Integer.valueOf(tmp.get(3)))){
|
|
proven = true;
|
|
}
|
|
}
|
|
else if(alphaCDR3toCDR1Map.get(Integer.valueOf(tmp.get(0))).equals(Integer.valueOf(tmp.get(3)))){
|
|
if(betaCDR3toCDR1Map.get(Integer.valueOf(tmp.get(1))).equals(Integer.valueOf(tmp.get(2)))){
|
|
proven = true;
|
|
}
|
|
}
|
|
tmp.add(proven.toString());
|
|
allResults.add(tmp);
|
|
if(proven){
|
|
trueCount++;
|
|
}
|
|
}
|
|
|
|
comments = new ArrayList<>();
|
|
comments.add("Plate size: " + samplePlate.getSize() + " wells");
|
|
comments.add("Previous pairs found: " + previousMatches.size());
|
|
comments.add("High overlap threshold: " + highThreshold);
|
|
comments.add("Low overlap threshold: " + lowThreshold);
|
|
comments.add("CDR1 matches attempted: " + allResults.size());
|
|
attemptRate = (double) allResults.size() / previousMatches.size();
|
|
comments.add("Matching attempt rate: " + attemptRate);
|
|
comments.add("Number of correct matches: " + trueCount);
|
|
correctRate = (double) trueCount / allResults.size();
|
|
comments.add("Correct matching rate: " + correctRate);
|
|
comments.add("Simulation time: " + nf.format(time.toSeconds()) + " seconds");
|
|
|
|
for(String s: comments){
|
|
System.out.println(s);
|
|
}
|
|
|
|
System.out.println("Simulation time: " + nf.format(time.toSeconds()) + " seconds");
|
|
MatchingResult dualTest = new MatchingResult(samplePlate.getSourceFileName(), comments, headers,
|
|
allResults, dualMatchesMap, time);
|
|
MatchingResult[] output = {firstTest, dualTest};
|
|
return output;
|
|
}
|
|
|
|
|
|
//Counts the well occupancy of the row peptides and column peptides into given maps, and
|
|
//fills weights in the given 2D array
|
|
private static void countSequencesAndFillMatrix(Plate samplePlate,
|
|
Map<Integer,Integer> allRowSequences,
|
|
Map<Integer,Integer> allColumnSequences,
|
|
Map<Integer,Integer> rowSequenceToVertexMap,
|
|
Map<Integer,Integer> columnSequenceToVertexMap,
|
|
int[] rowSequenceIndices,
|
|
int[] colSequenceIndices,
|
|
Map<Integer, Integer> rowSequenceCounts,
|
|
Map<Integer,Integer> columnSequenceCounts,
|
|
double[][] weights){
|
|
Map<Integer, Integer> wellNRowPeptides = null;
|
|
Map<Integer, Integer> wellNColumnPeptides = null;
|
|
int vertexStartValue = rowSequenceToVertexMap.size();
|
|
int numWells = samplePlate.getSize();
|
|
for (int n = 0; n < numWells; n++) {
|
|
wellNRowPeptides = samplePlate.assayWellsPeptideP(n, rowSequenceIndices);
|
|
for (Integer a : wellNRowPeptides.keySet()) {
|
|
if(allRowSequences.containsKey(a)){
|
|
rowSequenceCounts.merge(a, 1, (oldValue, newValue) -> oldValue + newValue);
|
|
}
|
|
}
|
|
wellNColumnPeptides = samplePlate.assayWellsPeptideP(n, colSequenceIndices);
|
|
for (Integer b : wellNColumnPeptides.keySet()) {
|
|
if(allColumnSequences.containsKey(b)){
|
|
columnSequenceCounts.merge(b, 1, (oldValue, newValue) -> oldValue + newValue);
|
|
}
|
|
}
|
|
for (Integer i : wellNRowPeptides.keySet()) {
|
|
if(allRowSequences.containsKey(i)){
|
|
for (Integer j : wellNColumnPeptides.keySet()) {
|
|
if(allColumnSequences.containsKey(j)){
|
|
weights[rowSequenceToVertexMap.get(i)][columnSequenceToVertexMap.get(j) - vertexStartValue] += 1.0;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
}
|
|
}
|
|
|
|
|
|
private static void filterByOccupancyThreshold(SimpleWeightedGraph<Integer, DefaultWeightedEdge> graph,
|
|
int low, int high) {
|
|
for(DefaultWeightedEdge e: graph.edgeSet()){
|
|
if ((graph.getEdgeWeight(e) > high) || (graph.getEdgeWeight(e) < low)){
|
|
graph.setEdgeWeight(e, 0.0);
|
|
}
|
|
}
|
|
}
|
|
private static void filterByOccupancyThreshold(Map<Integer, Integer> wellMap, int low, int high){
|
|
List<Integer> noise = new ArrayList<>();
|
|
for(Integer k: wellMap.keySet()){
|
|
if((wellMap.get(k) > high) || (wellMap.get(k) < low)){
|
|
noise.add(k);
|
|
}
|
|
}
|
|
for(Integer k: noise) {
|
|
wellMap.remove(k);
|
|
}
|
|
}
|
|
|
|
//Remove edges for pairs with large occupancy discrepancy
|
|
private static void filterByRelativeOccupancy(SimpleWeightedGraph<Integer, DefaultWeightedEdge> graph,
|
|
Map<Integer, Integer> alphaWellCounts,
|
|
Map<Integer, Integer> betaWellCounts,
|
|
Map<Integer, Integer> plateVtoAMap,
|
|
Map<Integer, Integer> plateVtoBMap,
|
|
Integer maxOccupancyDifference) {
|
|
for (DefaultWeightedEdge e : graph.edgeSet()) {
|
|
Integer alphaOcc = alphaWellCounts.get(plateVtoAMap.get(graph.getEdgeSource(e)));
|
|
Integer betaOcc = betaWellCounts.get(plateVtoBMap.get(graph.getEdgeTarget(e)));
|
|
//Adjust this to something cleverer later
|
|
if (Math.abs(alphaOcc - betaOcc) >= maxOccupancyDifference) {
|
|
graph.setEdgeWeight(e, 0.0);
|
|
}
|
|
}
|
|
}
|
|
|
|
//Remove edges for pairs where overlap size is significantly lower than the well occupancy
|
|
private static void filterByOverlapSize(SimpleWeightedGraph<Integer, DefaultWeightedEdge> graph,
|
|
Map<Integer, Integer> alphaWellCounts,
|
|
Map<Integer, Integer> betaWellCounts,
|
|
Map<Integer, Integer> plateVtoAMap,
|
|
Map<Integer, Integer> plateVtoBMap,
|
|
Integer minOverlapPercent) {
|
|
for (DefaultWeightedEdge e : graph.edgeSet()) {
|
|
Integer alphaOcc = alphaWellCounts.get(plateVtoAMap.get(graph.getEdgeSource(e)));
|
|
Integer betaOcc = betaWellCounts.get(plateVtoBMap.get(graph.getEdgeTarget(e)));
|
|
double weight = graph.getEdgeWeight(e);
|
|
double min = minOverlapPercent / 100.0;
|
|
if ((weight / alphaOcc < min) || (weight / betaOcc < min)) {
|
|
graph.setEdgeWeight(e, 0.0);
|
|
}
|
|
}
|
|
}
|
|
|
|
private static Map<Integer, Integer> makeSequenceToSequenceMap(List<Integer[]> cells, int keySequenceIndex,
|
|
int valueSequenceIndex){
|
|
Map<Integer, Integer> keySequenceToValueSequenceMap = new HashMap<>();
|
|
for (Integer[] cell : cells) {
|
|
keySequenceToValueSequenceMap.put(cell[keySequenceIndex], cell[valueSequenceIndex]);
|
|
}
|
|
return keySequenceToValueSequenceMap;
|
|
}
|
|
|
|
private static Map<Integer, Integer> makeVertexToSequenceMap(Map<Integer, Integer> sequences, Integer startValue) {
|
|
Map<Integer, Integer> map = new LinkedHashMap<>(); //LinkedHashMap to preserve order of entry
|
|
Integer index = startValue;
|
|
for (Integer k: sequences.keySet()) {
|
|
map.put(index, k);
|
|
index++;
|
|
}
|
|
return map;
|
|
}
|
|
|
|
private static Map<Integer, Integer> invertVertexMap(Map<Integer, Integer> map) {
|
|
Map<Integer, Integer> inverse = new HashMap<>();
|
|
for (Integer k : map.keySet()) {
|
|
inverse.put(map.get(k), k);
|
|
}
|
|
return inverse;
|
|
}
|
|
|
|
}
|