]> git.donarmstrong.com Git - mothur.git/blob - chimera.cpp
finished with ccode, returned bellerophon to last save before move, cleaned up pintai...
[mothur.git] / chimera.cpp
1 /*
2  *  chimera.cpp
3  *  Mothur
4  *
5  *  Created by Sarah Westcott on 8/11/09.
6  *  Copyright 2009 Schloss Lab Umass Amherst. All rights reserved.
7  *
8  */
9
10 #include "chimera.h"
11
12 //***************************************************************************************************************
13 //this is a vertical filter
14 void Chimera::createFilter(vector<Sequence*> seqs) {
15         try {
16                 
17                 int threshold = int (0.5 * seqs.size());
18 //cout << "threshhold = " << threshold << endl;
19                 
20                 vector<int> gaps;       gaps.resize(seqs[0]->getAligned().length(), 0);
21                 vector<int> a;          a.resize(seqs[0]->getAligned().length(), 0);
22                 vector<int> t;          t.resize(seqs[0]->getAligned().length(), 0);
23                 vector<int> g;          g.resize(seqs[0]->getAligned().length(), 0);
24                 vector<int> c;          c.resize(seqs[0]->getAligned().length(), 0);
25                 
26                 filterString = (string(seqs[0]->getAligned().length(), '1'));
27                 
28                 //for each sequence
29                 for (int i = 0; i < seqs.size(); i++) {
30                 
31                         string seqAligned = seqs[i]->getAligned();
32                         
33                         for (int j = 0; j < seqAligned.length(); j++) {
34                                 //if this spot is a gap
35                                 if ((seqAligned[j] == '-') || (seqAligned[j] == '.'))   {       gaps[j]++;      }
36                                 else if (toupper(seqAligned[j]) == 'A')                                 {       a[j]++;         }
37                                 else if (toupper(seqAligned[j]) == 'T')                                 {       t[j]++;         }
38                                 else if (toupper(seqAligned[j]) == 'G')                                 {       g[j]++;         }
39                                 else if (toupper(seqAligned[j]) == 'C')                                 {       c[j]++;         }
40                         }
41                 }
42                 
43                 //zero out spot where all sequences have blanks
44                 int numColRemoved = 0;
45                 for(int i = 0;i < seqs[0]->getAligned().length(); i++){
46                         if(gaps[i] == seqs.size())      {       filterString[i] = '0';  numColRemoved++;  }
47                         
48                         else if (((a[i] < threshold) && (t[i] < threshold) && (g[i] < threshold) && (c[i] < threshold))) {      filterString[i] = '0';  numColRemoved++;  }
49                         //cout << "a = " << a[i] <<  " t = " << t[i] <<  " g = " << g[i] <<  " c = " << c[i] << endl;
50                 }
51 //cout << "filter = " << filterString << endl;  
52
53                 mothurOut("Filter removed " + toString(numColRemoved) + " columns.");  mothurOutEndLine();
54         }
55         catch(exception& e) {
56                 errorOut(e, "Chimera", "createFilter");
57                 exit(1);
58         }
59 }
60 //***************************************************************************************************************
61 void Chimera::runFilter(vector<Sequence*> seqs) {
62         try {
63                 
64                 //for each sequence
65                 for (int i = 0; i < seqs.size(); i++) {
66                 
67                         string seqAligned = seqs[i]->getAligned();
68                         string newAligned = "";
69                         
70                         for (int j = 0; j < seqAligned.length(); j++) {
71                                 //if this spot is a gap
72                                 if (filterString[j] == '1') { newAligned += seqAligned[j]; }
73                         }
74                         
75                         seqs[i]->setAligned(newAligned);
76                 }
77                 
78         }
79         catch(exception& e) {
80                 errorOut(e, "Chimera", "runFilter");
81                 exit(1);
82         }
83 }
84
85 //***************************************************************************************************************
86 vector<Sequence*> Chimera::readSeqs(string file) {
87         try {
88                 ifstream in;
89                 openInputFile(file, in);
90                 vector<Sequence*> container;
91                 
92                 //read in seqs and store in vector
93                 while(!in.eof()){
94                         
95                         Sequence* current = new Sequence(in);
96                         container.push_back(current);
97                         gobble(in);
98                 }
99                 
100                 in.close();
101                 return container;
102         }
103         catch(exception& e) {
104                 errorOut(e, "Chimera", "readSeqs");
105                 exit(1);
106         }
107 }
108 //***************************************************************************************************************
109 void Chimera::setMask(string filename) {
110         try {
111                 
112                 if (filename == "default") {
113                         //default is from wigeon  236627 EU009184.1 Shigella dysenteriae str. FBD013
114                         seqMask = ".....................................................................................................AAATTGAAGAGTTT-GA--T-CA-T-G-GCTC-AG-AT-TGAA-C-GC--TGG-C--G-GC-A-GG--C----C-T--AACACA-T-GC-A-AGT-CGA-A-CG----------G-TAA-CA-G----------------------------GAAG-A-AG----------------------------------------------------CTT-G----------------------------------------------------------------------------------CT-TCTTT----------------G-CT--G--AC--G--AG-T-GG-C-GG-A--C-------------GGG-TGAGT-A--AT-GT-C-T-G-GG---A-A--A-CT-G--C-C-TGA--TG-G------------------------------------------------------------------A-GG----GGG-AT-AA-CTA-------------------------C-T-G-----------------------GAA-A---CGG-TAG-CTAA-TA---CC-G--C-AT-A----------A--------------------C-------------------------------------GT-C-----------------------------------------------------------------------------------------------------------------------G-CA-A--------------------------------------------------------------------------------------------------------------------------------------G-A-C---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------CAAA--G-A-G-GG-----G--GA-C-CT--------------------------------------------------------------------------------------------------------------------TCG-G----------------------------------------------------------------------------------------------------------------------G----CC-TC--T---T-G--------------C----C-A---T-CG-G---AT---G-T-----G-CCC-AGA--T-GGG--A------TT--A--G-CT-A----G---TAGG-T-G-GG-G-T----AAC-GG-C-T-C-ACCT--A-GG-C-G--A-CG-A------------TCC-C-T------AG-CT-G-G-TCT-G-AG----A--GG-AT--G-AC-C-AG-CCAC-A-CTGGA--A-C-TG-A-GA-C-AC-G-G-TCCAGA-CTCC-TAC-G--G-G-A-G-GC-A-GC-A-G-TG---GG-G-A-ATA-TTGCA-C-AA-T-GG--GC-GC-A----A-G-CC-T-GA-TG-CA-GCCA-TGCC-G-CG-T---G-T-A--T--GA-A-G--A--A-G-G-CC-----TT-CG---------G-G-T-T-G-T--A---AA-G-TAC--------TT-TC-A-G--C-GGG----GA-G--G---AA-GGGA---GTAA-AG----T--T--AA-T---A----C-----CT-T-TGC-TCA-TT-GA-CG-TT-A-C-CC-G-CA-G---------AA-----------GAAGC-ACC-GG-C-TAA---C--T-CCGT--GCCA--G-C---A--GCCG---C-GG--TA-AT--AC---GG-AG-GGT-GCA-A-G-CG-TTAA-T-CGG-AA-TT-A--C-T--GGGC-GTA----AA-GCGC-AC--G-CA-G-G-C-G------------G--T-TT-G-T-T-AA----G-T-C-A---G-ATG-TG-A-AA-TC--CC-CGG-G--------------------------------------------------------------------CT-C-AA-------------------------------------------------------------------------CC-T-G-GG-AA-C----T-G-C-A-T-C--------T--GA-T-A-C-T-G-GCA--A-G-C---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------T-T-G-A-G-T-C-----T-CG--TA-G-A------------G-GG-G-GG-T----AG--AATT-CCA-G-GT--GT-A-GCG-GTGAAA-TG-CGT-AGAG-A-TC-T-GGA--GG-A-AT-A-CC-GG--T--G--GC-GAA-G--G-C---G----G--C-C-CCCTG------G-AC-GA--------------------------------------------------------------AG-A-C-T--GA--CG-----CT-CA-GG--T-G-CGA--AA-G-C--------------G-TGGG-GAG-C-A-AACA--GG-ATTA-G-ATA-C-----CC-T-G-GTA-G-T----C-CA--C-G-CCG-T-AAA--C-GATG-TC--GA-CT---------T-GG--A--G-G-TT-G-TG-C--C--------------------------------------------------------------------------------------CTT-GA--------------------------------------------------------------------------------------------------------------------------------------------------G-G-C-GT--G-G-C-T-TC-C------GG--A----GC-TAA--CG-C-G-T--T--AA-GT--C----G-ACC-GCC-T-G-GG-GAG-TA---CGG-----C-C--G-C-A-A-GGT-T--AAA-ACTC-AAA---------TGAA-TTG-ACGGG-G-G-CCCG----C-A--C-A-A-GCG-GT-G--G--AG-CA-T--GT-GGT-TT-AATT-C-G-ATG-CAAC-G-CG-A-AG-A-A-CC-TT-A-CC-TGGTC-TT-G-AC-A-T-C--------------CAC-G-G-------------A-AG-T-T-T--TC--A-GA-G-A-T--G-A-G--A-A-T-G--T-G-----CC-------------------------------------T--TC-G------------------------------------------GG----A----A---CC-GTG---A--GA---------------------------------------------------C-A-G-G-T-GCTG-CA-TGG-CT--GTC-GTC-A-GC-TC---G-TG-TT-G--TGA-AA-TGT-T-GG-G-TT-AA-GT-CCCGC-AA--------C-GAG-CGC-A-ACC-C-T-TA--TC--C-TTTG--T-T-G-C-C---AG-C-G-----G-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------TCC------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------GG---C----C-G------------G----G---A-A--CT---------------C-A-A-A-G-GA-G--AC-T-G-CCA--G-T------------------------------------G-A---TAA----------------------------------A-C-T-G--G-A-GG-A--AGG-T--GGGG-A-TGAC-GTC--AAGT-C---ATC-A-T-G-G-C-C-CTT----AC-G--AC-C-A-GG-GC-TA-CAC-ACGTG-C--TA--CAATG---G-CGCA-T-A--C-AAA-GA-GA--------------------------------------------------------------------------------------------------A-G-C-G-A--C-CTCG-C--G---------------------------------------A-GA-G-C-----------A--A-G-CG---G----------A--CCT-C------A-T-AAAGT-GC-G-T-C-G-TAG-TCC--------GGA-T-TGGAG-TC--T-GCAA-CT-C-------------------------------------------------------------------------------------------------G-ACTCC-A-T-G-AA-G-TC-GGAAT-CG-C-TA--G-TA-AT-C-G-T----GGA-TC-A-G--A------AT--GCC-AC-G-GT-G-AAT-ACGT-T-CCCGGGCCT-TGTA----CACACCG-CCC-GTC-----A---CA--CCA-TG-GG-A--G---TGG-G-TT-GC-AAA--A-GAA------G--T-AGG-TA-G-C-T-T-AA-C-C--------------------------------------------------------------TT----C-------------------------------------------------------------------------------------------------G--GG-A--GG-G--C---GC-TTA--CC--ACT-T----T-GTG-AT-TCA------------------------TG--ACT-GGGG-TG-AAG-TCGTAACAA-GGTAA-CCGT-AGGGGAA-CCTG-CGGT-TGGATCACCTCCTTA................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................";
115                 }else if (filename == "") {  //do nothing 
116                         seqMask = "";
117                 }else{
118                         ifstream infile;
119                         openInputFile(filename, infile);
120                         
121                         while (!infile.eof()) {
122                                 Sequence temp(infile);
123                                 seqMask = temp.getAligned();
124                                 
125                                 gobble(infile);
126                         }
127                         
128                         infile.close();
129                 }
130         }
131         catch(exception& e) {
132                 errorOut(e, "Chimera", "setMask");
133                 exit(1);
134         }
135 }
136 //***************************************************************************************************************
137
138 vector< vector<float> > Chimera::readQuantiles() {
139         try {
140         
141                 ifstream in;
142                 openInputFile(quanfile, in);
143                 
144                 vector< vector<float> > quan;
145                 vector <float> temp;
146                 
147                 //to fill 0
148                 quan.push_back(temp); 
149         
150                 int num; float ten, twentyfive, fifty, seventyfive, ninetyfive, ninetynine; 
151                 
152                 while(!in.eof()){
153                         
154                         in >> num >> ten >> twentyfive >> fifty >> seventyfive >> ninetyfive >> ninetynine; 
155                         
156                         temp.clear();
157                         
158                         temp.push_back(ten); 
159                         temp.push_back(twentyfive);
160                         temp.push_back(fifty);
161                         temp.push_back(seventyfive);
162                         temp.push_back(ninetyfive);
163                         temp.push_back(ninetynine);
164                         
165                         quan.push_back(temp);  
166         
167                         gobble(in);
168                 }
169                 
170                 in.close();
171                 return quan;
172                 
173         }
174         catch(exception& e) {
175                 errorOut(e, "Chimera", "readQuantiles");
176                 exit(1);
177         }
178 }
179 //***************************************************************************************************************
180
181
182
183