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
|
package bin;
import java.util.Arrays;
import shared.Parse;
import shared.Tools;
/** Mostly written by ChatGPT and modified by me */
public class SimilarityMeasures {
public static void main(String[] args) {
float[] sample1={0.1f, 0.2f, 0.3f, 0.4f};
float[] sample2={0.1f, 0.2f, 0.4f, 0.3f};
int[] sample1i={1, 2, 3, 4};
int[] sample2i={1, 2, 4, 3};
int[] sample3i={2, 4, 6, 8};
int[] sample4i={8, 6, 4, 2};
// Print the similarity vector
System.out.println("Difference Vector Float12: "+Arrays.toString(calculateDifferenceVector(sample1, sample2)));
System.out.println("Difference Vector Int12: "+Arrays.toString(calculateDifferenceVector(sample1i, sample2i)));
System.out.println("Difference Vector Int13: "+Arrays.toString(calculateDifferenceVector(sample1i, sample3i)));
System.out.println("Difference Vector Int14: "+Arrays.toString(calculateDifferenceVector(sample1i, sample4i)));
}
public static boolean parse(String arg, String a, String b){
if(a.equals("null")){
//Do nothing
}else if(a.equals("cosine") || a.equals("cos")){
COSINE=Parse.parseBoolean(b);
}else if(a.equals("gccompensated")){
GC_COMPENSATED=Parse.parseBoolean(b);
}else if(a.equals("euclid") || a.equals("euc")){
EUCLID=Parse.parseBoolean(b);
}else if(a.equals("absolute") || a.equals("abs")){
ABSOLUTE=Parse.parseBoolean(b);
}else if(a.equals("jsd")){
JSD=Parse.parseBoolean(b);
}else if(a.equals("hellinger") || a.equals("hell") || a.equals("hel")){
HELLINGER=Parse.parseBoolean(b);
}else if(a.equals("ks") || a.equals("kst")){
KST=Parse.parseBoolean(b);
}else {
return false;
}
return true;
}
public static float[] calculateDifferenceVector(float[] a, float[] b) {
// float cosineSimilarity=cosineSimilarity(a, b);
float cosineDifference=cosineDifference(a, b);
float euclideanDistance=euclideanDistance(a, b);
float absoluteDifference=absDif(a, b);
float jensenShannonDivergence=jensenShannonDivergence(a, b);
float hellingerDistance=hellingerDistance(a, b);
float ksDifference=ksTest(a, b);
return new float[] {
cosineDifference,
euclideanDistance,
absoluteDifference,
jensenShannonDivergence,
hellingerDistance,
ksDifference
};
}
//For setting thresholds before neural net is implemented
public static float calculateDifferenceAverage(int[] a, int[] b) {
float inva=1f/Tools.max(1, Tools.sum(a));
float invb=1f/Tools.max(1, Tools.sum(b));
float cosineDifference=(COSINE ? cosineDifference(a, b, inva, invb) : 0);
float euclideanDistance=(EUCLID ? euclideanDistance(a, b, inva, invb) : 0);
float absoluteDifference=(ABSOLUTE ? absDif(a, b, inva, invb) : 0);
float jensenShannonDivergence=(JSD ? jensenShannonDivergence(a, b, inva, invb) : 0);
float hellingerDistance=(HELLINGER? hellingerDistance(a, b, inva, invb) : 0);
float ksDifference=(KST ? ksTest(a, b, inva, invb) : 0);
int div=(COSINE ? 1 : 0)+(EUCLID ? 1 : 0)+(ABSOLUTE ? 1 : 0)+(JSD ? 1 : 0)+(HELLINGER ? 1 : 0)+(KST ? 1 : 0);
float ret=(cosineDifference+euclideanDistance+absoluteDifference+
jensenShannonDivergence+hellingerDistance+ksDifference)/div;
return (Float.isFinite(ret) && ret>0 ? ret : 0);
}
public static float[] calculateDifferenceVector(int[] a, int[] b) {
float inva=1f/Tools.max(1, Tools.sum(a));
float invb=1f/Tools.max(1, Tools.sum(b));
// float cosineSimilarity=cosineSimilarity(a, b, inva, invb);
float cosineDifference=cosineDifference(a, b, inva, invb);
float euclideanDistance=euclideanDistance(a, b, inva, invb);
float absoluteDifference=absDif(a, b, inva, invb);
float jensenShannonDivergence=jensenShannonDivergence(a, b, inva, invb);
float hellingerDistance=hellingerDistance(a, b, inva, invb);
float ksDifference=ksTest(a, b, inva, invb);
return new float[] {
cosineDifference,
euclideanDistance,
absoluteDifference,
jensenShannonDivergence,
hellingerDistance,
ksDifference
};
}
public static float cosineDifference(float[] a, float[] b) {
return 1-cosineSimilarity(a, b);
}
public static float cosineSimilarity(float[] a, float[] b) {
float dotProduct=0f;
float normVec1=0f;
float normVec2=0f;
for (int i=0; i<a.length; i++) {
float ai=a[i], bi=b[i];
dotProduct+=ai*bi;
normVec1+=ai*ai;
normVec2+=bi*bi;
}
float ret=(float)(dotProduct/(Math.sqrt(normVec1)*Math.sqrt(normVec2)));
return (Float.isFinite(ret) && ret>0 ? ret : 0);
}
public static float cosineDifference(int[] a, int[] b) {
float inva=1f/Tools.max(1, Tools.sum(a));
float invb=1f/Tools.max(1, Tools.sum(b));
float ret=1-cosineSimilarity(a, b, inva, invb);
return (Float.isFinite(ret) && ret>0 ? ret : 0);
}
public static float cosineDifference(int[] a, int[] b, float inva, float invb) {
return 1-cosineSimilarity(a, b, inva, invb);
}
public static float cosineDifferenceCompensated(int[] a, int[] b, int k) {
return 1-cosineSimilarityCompensated(a, b, k, BinObject.gcmapMatrix[k]);
}
public static float cosineSimilarity(int[] a, int[] b, float inva, float invb) {
if(GC_COMPENSATED) {return cosineSimilarityCompensated(a, b, 4, BinObject.gcmapMatrix[4]);}
float dotProduct=0f;
float normVec1=0f;
float normVec2=0f;
for (int i=0; i<a.length; i++) {
float ai=a[i]*inva, bi=b[i]*invb;
dotProduct+=ai*bi;
normVec1+=ai*ai;
normVec2+=bi*bi;
}
return (float)(dotProduct/(Math.sqrt(normVec1)*Math.sqrt(normVec2)));
}
public static float cosineSimilarityCompensated(int[] a, int[] b, int k, int[] gcmap) {
float[] aSum=new float[k+1];
float[] bSum=new float[k+1];
for(int i=0; i<a.length; i++) {
int gc=gcmap[i];
aSum[gc]+=a[i];
bSum[gc]+=b[i];
}
for(int i=0; i<aSum.length; i++) {
aSum[i]=1f/Math.max(aSum[i], 1);
bSum[i]=1f/Math.max(bSum[i], 1);
}
float dotProduct=0f;
float normVec1=0f;
float normVec2=0f;
for (int i=0; i<a.length; i++) {
int gc=gcmap[i];
float ai=a[i]*aSum[gc], bi=b[i]*bSum[gc];
dotProduct+=ai*bi;
normVec1+=ai*ai;
normVec2+=bi*bi;
}
return (float)(dotProduct/(Math.sqrt(normVec1)*Math.sqrt(normVec2)));
}
public static float euclideanDistance(float[] a, float[] b) {
float sumSquaredDifferences=0f;
for (int i=0; i<a.length; i++) {
float ai=a[i], bi=b[i];
float d=ai-bi;
sumSquaredDifferences+=d*d;
}
return (float)Math.sqrt(sumSquaredDifferences);
}
public static float euclideanDistance(int[] a, int[] b) {
float inva=1f/Tools.max(1, Tools.sum(a));
float invb=1f/Tools.max(1, Tools.sum(b));
float ret=euclideanDistance(a, b, inva, invb);
return (Float.isFinite(ret) && ret>0 ? ret : 0);
}
public static float euclideanDistance(int[] a, int[] b, float inva, float invb) {
float sumSquaredDifferences=0f;
for (int i=0; i<a.length; i++) {
float ai=a[i]*inva, bi=b[i]*invb;
float d=ai-bi;
sumSquaredDifferences+=d*d;
}
return (float)Math.sqrt(sumSquaredDifferences);
}
/**
* @param a Contig kmer frequencies
* @param b Cluster kmer frequencies
* @return Score
*/
static final float absDif(float[] a, float[] b){
assert(a.length==b.length);
double sum=0;
for(int i=0; i<a.length; i++){
sum+=Math.abs(a[i]-b[i]);
}
return (float)sum;
}
/**
* @param a Contig kmer counts
* @param b Cluster kmer counts
* @return Score
*/
static final float absDif(int[] a, int[] b, float inva, float invb){
assert(a.length==b.length);
float sum=0;
for(int i=0; i<a.length; i++){
float ai=a[i]*inva, bi=b[i]*invb;
sum+=Math.abs(ai-bi);
}
return sum;
}
public static float jensenShannonDivergence(float[] a, float[] b) {
float kldSumA=0, kldSumB=0;
for (int i=0; i<a.length; i++) {
float ai=a[i]+0.0005f, bi=b[i]+0.0005f;//Prevents zero values
float avgi=(ai+bi)*0.5f;
kldSumA+=ai*Math.log(ai/avgi);
kldSumA+=bi*Math.log(bi/avgi);
}
return (kldSumA+kldSumB)*invLog2*0.5f;
}
public static float jensenShannonDivergence(int[] a, int[] b, float inva, float invb) {
float kldSumA=0, kldSumB=0;
for (int i=0; i<a.length; i++) {
float ai=a[i]*inva+0.0005f, bi=b[i]*invb+0.0005f;//Prevents zero values
float avgi=(ai+bi)*0.5f;
kldSumA+=ai*Math.log(ai/avgi);
kldSumA+=bi*Math.log(bi/avgi);
}
return (kldSumA+kldSumB)*invLog2*0.5f;
}
// public static float jensenShannonDivergenceSlow(float[] a, float[] b) {
// float[] avg=new float[a.length];
// for (int i=0; i<a.length; i++) {
// float ai=a[i], bi=b[i];
// avg[i]=(ai+bi)*0.5f;
// }
//
// return (kullbackLeiblerDivergence(a, avg)+kullbackLeiblerDivergence(b, avg))*0.5f;
// }
//
// public static float kullbackLeiblerDivergence(float[] p, float[] q) {
// float sum=0f;
// for (int i=0; i<p.length; i++) {
// float pi=p[i], qi=q[i];
// if (p[i]!=0) {
// sum+=p[i]*Math.log(pi/qi);
// }
// }
// return sum*invLog2;
// }
public static float hellingerDistance(float[] a, float[] b) {
float sum=0f;
for (int i=0; i<a.length; i++) {
float ai=a[i], bi=b[i];
float d=(float)(Math.sqrt(ai)-Math.sqrt(bi));
sum+=d*d;
}
return (float)Math.sqrt(sum)*invRoot2;
}
public static float hellingerDistance(int[] a, int[] b, float inva, float invb) {
float sum=0f;
for (int i=0; i<a.length; i++) {
float ai=a[i]*inva, bi=b[i]*invb;
float d=(float)(Math.sqrt(ai)-Math.sqrt(bi));
sum+=d*d;
}
return (float)Math.sqrt(sum)*invRoot2;
}
/** This is a KS test for binned histograms, not raw values */
public static float ksTest(float[] histogram1, float[] histogram2) {
// Ensure both histograms have the same length
if (histogram1.length!=histogram2.length) {
throw new IllegalArgumentException("Histograms must have the same number of bins");
}
float cd1=0, cd2=0, dMax=0;
// Compute the KS statistic (maximum absolute difference between the two CDFs)
for (int i=0; i<histogram1.length; i++) {
cd1+=histogram1[i];
cd2+=histogram2[i];
dMax=(float)Math.max(dMax, Math.abs(cd1-cd2));
}
return dMax;
}
/** This is a KS test for binned histograms, not raw values */
public static float ksTest(int[] a, int[] b, float inva, float invb) {
// Ensure both histograms have the same length
if (a.length!=b.length) {
throw new IllegalArgumentException("Histograms must have the same number of bins");
}
float cda=0, cdb=0, dMax=0;
// Compute the KS statistic (maximum absolute difference between the two CDFs)
for (int i=0; i<a.length; i++) {
float ai=a[i]*inva, bi=b[i]*invb;
cda+=ai;
cdb+=bi;
dMax=(float)Math.max(dMax, Math.abs(cda-cdb));
}
return dMax;
}
private static final float root2=(float)Math.sqrt(2);
private static final float log2=(float)Math.log(2);
private static final float invRoot2=1/root2;
private static final float invLog2=1/log2;
public static boolean GC_COMPENSATED=false;
//2531 kcps (times include contig loading)
//26 clusters
// Completeness Score: 60.278
// Contamination Score: 2.1925
public static boolean COSINE=true;
//2796 kcps
//21 clusters
//Completeness Score: 60.909
//Contamination Score: 2.3108
public static boolean EUCLID=false;//0.008
//2636 kcps
//23 clusters at 4x threshold of cosine
// Completeness Score: 60.947
// Contamination Score: 1.7679
public static boolean ABSOLUTE=false; //Best at 0.089
//183 kcps
//22 clusters
// Completeness Score: 59.169
// Contamination Score: 2.1959
public static boolean JSD=false;
//953 kcps
//~22 at 2x threshold of cosine
// Completeness Score: 61.072
// Contamination Score: 1.9358
public static boolean HELLINGER=false;//0.0425
//1859 kcps
//20 clusters
// Completeness Score: 26.380
// Contamination Score: 3.1930
public static boolean KST=false;
}
|
