int main(int argc, char **argv)
{
AjPSeqall seq1;
AjPSeqset seq2;
AjPSeq a;
const AjPSeq b;
AjPStr m = 0;
AjPStr n = 0;
AjPFile errorf;
AjBool show = ajFalse;
ajint lena = 0;
ajint lenb = 0;
const char *p;
const char *q;
AjPMatrixf matrix;
AjPSeqCvt cvt = 0;
float **sub;
ajint *compass = NULL;
float *path = NULL;
float gapopen;
float gapextend;
float score;
ajint begina;
ajint i;
ajuint k;
ajint beginb;
ajint start1 = 0;
ajint start2 = 0;
ajint end1 = 0;
ajint end2 = 0;
ajint width = 0;
AjPTable seq1MatchTable = 0;
ajint wordlen = 6;
ajint oldmax = 0;
AjPAlign align = NULL;
embInit("supermatcher", argc, argv);
matrix = ajAcdGetMatrixf("datafile");
seq1 = ajAcdGetSeqall("asequence");
seq2 = ajAcdGetSeqset("bsequence");
gapopen = ajAcdGetFloat("gapopen");
gapextend = ajAcdGetFloat("gapextend");
wordlen = ajAcdGetInt("wordlen");
align = ajAcdGetAlign("outfile");
errorf = ajAcdGetOutfile("errorfile");
width = ajAcdGetInt("width"); /* not the same as awidth */
gapopen = ajRoundFloat(gapopen, 8);
gapextend = ajRoundFloat(gapextend, 8);
sub = ajMatrixfGetMatrix(matrix);
cvt = ajMatrixfGetCvt(matrix);
embWordLength(wordlen);
ajSeqsetTrim(seq2);
while(ajSeqallNext(seq1,&a))
{
ajSeqTrim(a);
begina = 1 + ajSeqGetOffset(a);
m = ajStrNewRes(1+ajSeqGetLen(a));
lena = ajSeqGetLen(a);
ajDebug("Read '%S'\n", ajSeqGetNameS(a));
if(!embWordGetTable(&seq1MatchTable, a)) /* get table of words */
ajErr("Could not generate table for %s\n",
ajSeqGetNameC(a));
for(k=0;k<ajSeqsetGetSize(seq2);k++)
{
b = ajSeqsetGetseqSeq(seq2, k);
lenb = ajSeqGetLen(b);
beginb = 1 + ajSeqGetOffset(b);
ajDebug("Processing '%S'\n", ajSeqGetNameS(b));
p = ajSeqGetSeqC(a);
q = ajSeqGetSeqC(b);
if(!supermatcher_findstartpoints(seq1MatchTable,b,a,
&start1, &start2,
&end1, &end2))
{
ajFmtPrintF(errorf,
"No wordmatch start points for "
"%s vs %s. No alignment\n",
ajSeqGetNameC(a),ajSeqGetNameC(b));
continue;
}
n=ajStrNewRes(1+ajSeqGetLen(b));
ajStrAssignC(&m,"");
ajStrAssignC(&n,"");
ajDebug("++ %S v %S start:%d %d end:%d %d\n",
ajSeqGetNameS(a), ajSeqGetNameS(b),
start1, start2, end1, end2);
if(end1-start1+1 > oldmax)
{
oldmax = ((end1-start1)+1);
AJRESIZE(path,oldmax*width*sizeof(float));
AJRESIZE(compass,oldmax*width*sizeof(ajint));
ajDebug("++ resize to oldmax: %d\n", oldmax);
}
for(i=0;i<((end1-start1)+1)*width;i++)
path[i] = 0.0;
ajDebug("Calling embAlignPathCalcFast "
"%d..%d [%d/%d] %d..%d [%d/%d]\n",
start1, end1, (end1 - start1 + 1), lena,
start2, end2, (end2 - start2 + 1), lenb);
score = embAlignPathCalcSWFast(&p[start1],&q[start2],
end1-start1+1,end2-start2+1,
0,width,
gapopen,gapextend,
path,sub,cvt,
compass,show);
embAlignWalkSWMatrixFast(path,compass,gapopen,gapextend,a,b,
&m,&n,end1-start1+1,end2-start2+1,
0,width,
&start1,&start2);
if(!ajAlignFormatShowsSequences(align))
{
ajAlignDefineCC(align, ajStrGetPtr(m),
ajStrGetPtr(n), ajSeqGetNameC(a),
ajSeqGetNameC(b));
ajAlignSetScoreR(align, score);
}
else
{
embAlignReportLocal(align, a, b,
m,n,start1,start2,
gapopen, gapextend,
score,matrix, begina, beginb);
}
ajAlignWrite(align);
ajAlignReset(align);
ajStrDel(&n);
}
embWordFreeTable(&seq1MatchTable); /* free table of words */
seq1MatchTable=0;
ajStrDel(&m);
}
if(!ajAlignFormatShowsSequences(align))
{
ajMatrixfDel(&matrix);
}
AJFREE(path);
AJFREE(compass);
ajAlignClose(align);
ajAlignDel(&align);
ajSeqallDel(&seq1);
ajSeqDel(&a);
ajSeqsetDel(&seq2);
ajFileClose(&errorf);
embExit();
return 0;
}
int main(int argc, char **argv)
{
AjPSeqall queryseqs;
AjPSeqset targetseqs;
AjPSeq queryseq;
const AjPSeq targetseq;
AjPStr queryaln = 0;
AjPStr targetaln = 0;
AjPFile errorf;
AjBool show = ajFalse;
const char *queryseqc;
const char *targetseqc;
AjPMatrixf matrix;
AjPSeqCvt cvt = 0;
float **sub;
ajint *compass = NULL;
float *path = NULL;
float gapopen;
float gapextend;
float score;
float minscore;
ajuint j, k;
ajint querystart = 0;
ajint targetstart = 0;
ajint queryend = 0;
ajint targetend = 0;
ajint width = 0;
AjPTable kmers = 0;
ajint wordlen = 6;
ajint oldmax = 0;
ajint newmax = 0;
ajuint ntargetseqs;
ajuint nkmers;
AjPAlign align = NULL;
EmbPWordMatch maxmatch; /* match with maximum score */
/* Cursors for the current sequence being scanned,
** i.e., until which location it was scanned.
** Separate cursor/location entries for each sequence in the seqset.
*/
ajuint* lastlocation;
EmbPWordRK* wordsw = NULL;
AjPList* matchlist = NULL;
embInit("supermatcher", argc, argv);
matrix = ajAcdGetMatrixf("datafile");
queryseqs = ajAcdGetSeqall("asequence");
targetseqs= ajAcdGetSeqset("bsequence");
gapopen = ajAcdGetFloat("gapopen");
gapextend = ajAcdGetFloat("gapextend");
wordlen = ajAcdGetInt("wordlen");
align = ajAcdGetAlign("outfile");
errorf = ajAcdGetOutfile("errorfile");
width = ajAcdGetInt("width"); /* width for banded Smith-Waterman */
minscore = ajAcdGetFloat("minscore");
gapopen = ajRoundFloat(gapopen, 8);
gapextend = ajRoundFloat(gapextend, 8);
sub = ajMatrixfGetMatrix(matrix);
cvt = ajMatrixfGetCvt(matrix);
embWordLength(wordlen);
/* seqset sequence is the reference sequence for SAM format */
ajAlignSetRefSeqIndx(align, 1);
ajSeqsetTrim(targetseqs);
ntargetseqs = ajSeqsetGetSize(targetseqs);
AJCNEW0(matchlist, ntargetseqs);
/* get tables of words */
for(k=0;k<ntargetseqs;k++)
{
targetseq = ajSeqsetGetseqSeq(targetseqs, k);
embWordGetTable(&kmers, targetseq);
ajDebug("Number of distinct kmers found so far: %d\n",
ajTableGetLength(kmers));
}
AJCNEW0(lastlocation, ntargetseqs);
if(ajTableGetLength(kmers)<1)
ajErr("no kmers found");
nkmers = embWordRabinKarpInit(kmers, &wordsw, wordlen, targetseqs);
while(ajSeqallNext(queryseqs,&queryseq))
{
ajSeqTrim(queryseq);
queryaln = ajStrNewRes(1+ajSeqGetLen(queryseq));
ajDebug("Read '%S'\n", ajSeqGetNameS(queryseq));
for(k=0;k<ntargetseqs;k++)
{
lastlocation[k]=0;
matchlist[k] = ajListstrNew();
}
embWordRabinKarpSearch(ajSeqGetSeqS(queryseq), targetseqs,
(const EmbPWordRK*)wordsw, wordlen, nkmers,
matchlist, lastlocation, ajFalse);
for(k=0;k<ajSeqsetGetSize(targetseqs);k++)
{
targetseq = ajSeqsetGetseqSeq(targetseqs, k);
ajDebug("Processing '%S'\n", ajSeqGetNameS(targetseq));
if(ajListGetLength(matchlist[k])==0)
{
ajFmtPrintF(errorf,
"No wordmatch start points for "
"%s vs %s. No alignment\n",
ajSeqGetNameC(queryseq),ajSeqGetNameC(targetseq));
embWordMatchListDelete(&matchlist[k]);
continue;
}
/* only the maximum match is used as seed
* (if there is more than one location with the maximum match
* only the first one is used)
* TODO: we should add a new option to make above limit optional
*/
maxmatch = embWordMatchFirstMax(matchlist[k]);
supermatcher_findendpoints(maxmatch,targetseq, queryseq,
&targetstart, &querystart,
&targetend, &queryend);
targetaln=ajStrNewRes(1+ajSeqGetLen(targetseq));
queryseqc = ajSeqGetSeqC(queryseq);
targetseqc = ajSeqGetSeqC(targetseq);
ajStrAssignC(&queryaln,"");
ajStrAssignC(&targetaln,"");
ajDebug("++ %S v %S start:%d %d end:%d %d\n",
ajSeqGetNameS(targetseq), ajSeqGetNameS(queryseq),
targetstart, querystart, targetend, queryend);
newmax = (targetend-targetstart+2)*width;
if(newmax > oldmax)
{
AJCRESIZE0(path,oldmax,newmax);
AJCRESIZE0(compass,oldmax,newmax);
oldmax=newmax;
ajDebug("++ memory re/allocation for path/compass arrays"
" to size: %d\n", newmax);
}
else
{
AJCSET0(path,newmax);
AJCSET0(compass,newmax);
}
ajDebug("Calling embAlignPathCalcSWFast "
"%d..%d [%d/%d] %d..%d [%d/%d] width:%d\n",
querystart, queryend, (queryend - querystart + 1),
ajSeqGetLen(queryseq),
targetstart, targetend, (targetend - targetstart + 1),
ajSeqGetLen(targetseq),
width);
score = embAlignPathCalcSWFast(&targetseqc[targetstart],
&queryseqc[querystart],
targetend-targetstart+1,
queryend-querystart+1,
0,width,
gapopen,gapextend,
path,sub,cvt,
compass,show);
if(score>minscore)
{
embAlignWalkSWMatrixFast(path,compass,gapopen,gapextend,
targetseq,queryseq,
&targetaln,&queryaln,
targetend-targetstart+1,
queryend-querystart+1,
0,width,
&targetstart,&querystart);
if(!ajAlignFormatShowsSequences(align))
{
ajAlignDefineCC(align, ajStrGetPtr(targetaln),
ajStrGetPtr(queryaln),
ajSeqGetNameC(targetseq),
ajSeqGetNameC(queryseq));
ajAlignSetScoreR(align, score);
}
else
{
ajDebug(" queryaln:%S \ntargetaln:%S\n",
queryaln,targetaln);
embAlignReportLocal(align,
queryseq, targetseq,
queryaln, targetaln,
querystart, targetstart,
gapopen, gapextend,
score, matrix,
1 + ajSeqGetOffset(queryseq),
1 + ajSeqGetOffset(targetseq)
);
}
ajAlignWrite(align);
ajAlignReset(align);
}
ajStrDel(&targetaln);
embWordMatchListDelete(&matchlist[k]);
}
ajStrDel(&queryaln);
}
for(k=0;k<nkmers;k++)
{
AJFREE(wordsw[k]->seqindxs);
AJFREE(wordsw[k]->nSeqMatches);
for(j=0;j<wordsw[k]->nseqs;j++)
AJFREE(wordsw[k]->locs[j]);
AJFREE(wordsw[k]->nnseqlocs);
AJFREE(wordsw[k]->locs);
AJFREE(wordsw[k]);
}
embWordFreeTable(&kmers);
if(!ajAlignFormatShowsSequences(align))
ajMatrixfDel(&matrix);
AJFREE(path);
AJFREE(compass);
AJFREE(kmers);
AJFREE(wordsw);
AJFREE(matchlist);
AJFREE(lastlocation);
ajAlignClose(align);
ajAlignDel(&align);
ajSeqallDel(&queryseqs);
ajSeqDel(&queryseq);
ajSeqsetDel(&targetseqs);
ajFileClose(&errorf);
embExit();
return 0;
}
int main(int argc, char **argv)
{
ajint i;
ajint numseq;
ajint j = 0;
ajint numres;
ajint count;
ajint k;
ajint kmax;
float defheight;
float currentscale;
AjPStr shade = NULL;
AjPFloat pair = NULL;
AjPGraph graph = NULL;
AjPMatrix cmpmatrix = NULL;
AjPSeqCvt cvt = NULL;
AjPStr matcodes = NULL;
AjBool consensus;
AjBool colourbyconsensus;
AjBool colourbyresidues;
AjBool colourbyshade = AJFALSE;
AjBool boxit;
AjBool boxcol;
AjBool portrait;
AjBool collision;
ajint identity;
AjBool listoptions;
ajint alternative;
AjPStr altstr = NULL;
AjPStr sidentity = NULL;
AjPStr ssimilarity = NULL;
AjPStr sother = NULL;
AjPStr sboxcolval = NULL;
AjPStr options = NULL;
/* ajint showscore = 0; */
ajint iboxcolval = 0;
ajint cidentity = RED;
ajint csimilarity = GREEN;
ajint cother = BLACK;
float fxp;
float fyp;
float yincr;
float y;
ajint ixlen;
ajint iylen;
ajint ixoff;
ajint iyoff;
char res[2] = " ";
float *score = 0;
float scoremax = 0;
float *identical = NULL;
ajint identicalmaxindex;
float *matching = NULL;
ajint matchingmaxindex;
float *colcheck = NULL;
ajint **matrix;
ajint m1 = 0;
ajint m2 = 0;
ajint ms = 0;
ajint highindex = 0;
ajint myindex;
ajint *previous = 0;
AjBool iscons = ajFalse;
ajint currentstate = 0;
ajint oldfg = 0;
float fold = 0.0;
ajint *colmat = 0;
ajint *shadecolour = 0;
/* float identthresh = 1.5; */
/* float simthresh = 1.0; */
/* float relthresh = 0.5; */
float part = 0.0;
const char *cptr;
ajint resbreak;
float fplural;
float ystart;
float xmin;
float xmax;
float xmid;
AjPTime ajtime;
ajint gapcount = 0;
ajint countforgap = 0;
ajint boxindex;
float max;
ajint matsize;
ajint seqperpage = 0;
ajint startseq;
ajint endseq;
ajint newILend = 0;
ajint newILstart;
void *freeptr;
ajint itmp;
embInit("prettyplot", argc, argv);
seqset = ajAcdGetSeqset("sequences");
numres = ajAcdGetInt("residuesperline");
resbreak = ajAcdGetInt("resbreak");
ajSeqsetFill(seqset); /* Pads sequence set with gap characters */
numseq = ajSeqsetGetSize(seqset);
graph = ajAcdGetGraph("graph");
colourbyconsensus = ajAcdGetBoolean("ccolours");
colourbyresidues = ajAcdGetBoolean("docolour");
shade = ajAcdGetString("shade");
pair = ajAcdGetArray("pair");
identity = ajAcdGetInt("identity");
boxit = ajAcdGetBoolean("box");
ajtime = ajTimeNewTodayFmt("daytime");
ajSeqsetTrim(seqset);
/* offset = ajSeqsetGetOffset(seqset); Unused */
ajGraphAppendTitleS(graph, ajSeqsetGetUsa(seqset));
if(boxit)
{
AJCNEW(seqboxptr, numseq);
for(i=0;i<numseq;i++)
AJCNEW(seqboxptr[i], ajSeqsetGetLen(seqset));
}
boxcol = ajAcdGetBoolean("boxcol");
sboxcolval = ajAcdGetString("boxuse");
if(boxcol)
{
iboxcolval = ajGraphicsCheckColourS(sboxcolval);
if(iboxcolval == -1)
iboxcolval = GREY;
}
consensus = ajAcdGetBoolean("consensus");
if(consensus)
{
AJCNEW(constr, ajSeqsetGetLen(seqset)+1);
constr[0] = '\0';
}
shownames = ajAcdGetBoolean("name");
shownumbers = ajAcdGetBoolean("number");
charlen = ajAcdGetInt("maxnamelen");
fplural = ajAcdGetFloat("plurality");
portrait = ajAcdGetBoolean("portrait");
collision = ajAcdGetBoolean("collision");
listoptions = ajAcdGetBoolean("listoptions");
altstr = ajAcdGetListSingle("alternative");
cmpmatrix = ajAcdGetMatrix("matrixfile");
ajStrToInt(altstr, &alternative);
matrix = ajMatrixGetMatrix(cmpmatrix);
cvt = ajMatrixGetCvt(cmpmatrix);
matsize = ajMatrixGetSize(cmpmatrix);
AJCNEW(identical,matsize);
AJCNEW(matching,matsize);
AJCNEW(colcheck,matsize);
numgaps = numres/resbreak;
numgaps--;
if(portrait)
{
ajGraphicsSetPortrait(1);
ystart = (float) 75.0;
}
else
ystart = (float) 75.0;
/* pair is an array of three non-negative floats */
/* identthresh = ajFloatGet(pair,0); Unused */
/* simthresh = ajFloatGet(pair,1); Unused */
/* relthresh = ajFloatGet(pair,2); Unused */
/*
** shade is a formatted 4-character string. Characters BLPW only.
** controlled by a pattern in ACD.
*/
if(ajStrGetLen(shade))
{
AJCNEW(shadecolour,4);
cptr = ajStrGetPtr(shade);
for(i=0;i<4;i++){
if(cptr[i]== 'B' || cptr[i]== 'b')
shadecolour[i] = BLACK;
else if(cptr[i]== 'L' || cptr[i]== 'l')
shadecolour[i] = BROWN;
else if(cptr[i]== 'P' || cptr[i]== 'p')
shadecolour[i] = WHEAT;
else if(cptr[i]== 'W' || cptr[i]== 'w')
shadecolour[i] = WHITE;
}
colourbyconsensus = colourbyresidues = ajFalse;
colourbyshade = ajTrue;
}
/*
** we can colour by consensus or residue but not both
** if we have to choose, use the consensus
*/
if(colourbyconsensus && colourbyresidues)
colourbyconsensus = AJFALSE;
sidentity = ajAcdGetString("cidentity");
ssimilarity = ajAcdGetString("csimilarity");
sother = ajAcdGetString("cother");
if(colourbyconsensus)
{
cidentity = ajGraphicsCheckColourS(sidentity);
if(cidentity == -1)
cidentity = RED;
csimilarity = ajGraphicsCheckColourS(ssimilarity);
if(csimilarity == -1)
csimilarity = GREEN;
cother = ajGraphicsCheckColourS(sother);
if(cother == -1)
cother = BLACK;
}
else if(colourbyresidues)
{
matcodes = ajMatrixGetCodes(cmpmatrix);
if(ajSeqsetIsProt(seqset))
colmat = ajGraphicsBasecolourNewProt(matcodes);
else
colmat = ajGraphicsBasecolourNewNuc(matcodes);
}
/* output the options used as the subtitle for the bottom of the graph */
if(listoptions)
{
ajStrAssignC(&options,"");
ajFmtPrintAppS(&options,"-plurality %.1f",fplural);
if(collision)
ajStrAppendC(&options," -collision");
else
ajStrAppendC(&options," -nocollision");
if(boxit)
ajStrAppendC(&options," -box");
else
ajStrAppendC(&options," -nobox");
if(boxcol)
ajStrAppendC(&options," -boxcol");
else
ajStrAppendC(&options," -noboxcol");
if(colourbyconsensus)
ajStrAppendC(&options," -colbyconsensus");
else if(colourbyresidues)
ajStrAppendC(&options," -colbyresidues");
else if(colourbyshade)
ajStrAppendC(&options," -colbyshade");
else
ajStrAppendC(&options," -nocolour");
if(alternative==2)
ajStrAppendC(&options," -alt 2");
else if(alternative==1)
ajStrAppendC(&options," -alt 1");
else if(alternative==3)
ajStrAppendC(&options," -alt 3");
}
AJCNEW(seqcolptr, numseq);
for(i=0;i<numseq;i++)
AJCNEW(seqcolptr[i], ajSeqsetGetLen(seqset));
AJCNEW(seqcharptr, numseq);
AJCNEW(seqnames, numseq);
AJCNEW(score, numseq);
AJCNEW(previous, numseq);
AJCNEW(seqcount, numseq);
for(i=0;i<numseq;i++)
{
ajSeqsetFmtUpper(seqset);
seqcharptr[i] = ajSeqsetGetseqSeqC(seqset, i);
seqnames[i] = 0;
ajStrAppendS(&seqnames[i],ajSeqsetGetseqNameS(seqset, i));
ajStrTruncateLen(&seqnames[i],charlen);
previous[i] = 0;
seqcount[i] = 0;
}
/*
** user will pass the number of residues to fit a page
** therefore we now need to calculate the size of the chars
** based on this and get the new char width.
** 'charlen' maximum characters for the name (truncated above)
*/
ajGraphicsGetCharsize(&defheight,¤tscale);
xmin = -charlen - (float)2.0;
xmax = (float)numres+(float)11.0+(float)(numres/resbreak);
xmid = (xmax + xmin)/(float)2.0;
ajGraphOpenWin(graph, xmin, xmax,
(float)0.0, ystart+(float)1.0);
ajGraphGetParamsPage(graph, &fxp,&fyp,&ixlen,&iylen,&ixoff,&iyoff);
if(portrait)
{
itmp = ixlen;
ixlen = iylen;
iylen = itmp;
}
ajGraphicsGetCharsize(&defheight,¤tscale);
ajGraphicsSetCharscale(((float)ixlen/((float)(numres+charlen+1)*
(currentscale * (float) 1.5)))/
currentscale);
/* ajGraphicsSetCharscale(((float)ixlen/((float)(numres+charlen)*
(currentscale+(float)1.0)))/
currentscale); */
ajGraphicsGetCharsize(&defheight,¤tscale);
yincr = (currentscale + (float)3.0)*(float)0.3;
/*
** If we have titles (now the standard graph title and subtitle and footer)
** leave 7 rows of space for them
*/
y=ystart-(float)7.0;
if(ajStrGetLen(options))
{
fold = ajGraphicsSetCharscale(1.0);
ajGraphicsDrawposTextAtmid(xmid,2.0,
ajStrGetPtr(options));
ajGraphicsSetCharscale(fold);
}
/* if sequences per page not set then calculate it */
if(!seqperpage)
{
seqperpage = prettyplot_calcseqperpage(yincr,y,consensus);
if(seqperpage>numseq)
seqperpage=numseq;
}
count = 0;
/*
** for boxes we need to set a foreground colour for the box lines
** and save the current foreground colour
*/
if(boxit && boxcol)
oldfg = ajGraphicsSetFgcolour(iboxcolval);
/*
** step through each residue position
*/
kmax = ajSeqsetGetLen(seqset) - 1;
for(k=0; k<= kmax; k++)
{
/* reset column score array */
for(i=0;i<numseq;i++)
score[i] = 0.0;
/* reset matrix character testing arrays */
for(i=0;i<matsize;i++)
{
identical[i] = 0.0;
matching[i] = 0.0;
colcheck[i] = 0.0;
}
/* generate a score for this residue in each sequence */
for(i=0;i<numseq;i++)
{
m1 = ajSeqcvtGetCodeK(cvt, seqcharptr[i][k]);
for(j=0;j<numseq;j++)
{
m2 = ajSeqcvtGetCodeK(cvt, seqcharptr[j][k]);
if(m1 && m2)
score[i] += (float)matrix[m1][m2]*
ajSeqsetGetseqWeight(seqset, j);
}
if(m1)
identical[m1] += ajSeqsetGetseqWeight(seqset, i);
}
/* find the highest score */
highindex = -1;
scoremax = INT_MIN;
/*ajDebug("Scores at position %d:\n", k);*/
for(i=0;i<numseq;i++)
{
/*ajDebug(" seq %d: '%c' %f\n",i,seqcharptr[i][k],score[i]);*/
if(score[i] > scoremax)
{
scoremax = score[i];
highindex = i;
}
}
for(i=0;i<numseq;i++)
{
m1 = ajSeqcvtGetCodeK(cvt, seqcharptr[i][k]);
if(!matching[m1])
{
for(j=0;j<numseq;j++)
{
m2 = ajSeqcvtGetCodeK(cvt, seqcharptr[j][k]);
if(m1 && m2 && matrix[m1][m2] > 0)
matching[m1] += ajSeqsetGetseqWeight(seqset, j);
}
}
}
/* find highs for matching and identical */
matchingmaxindex = 0;
identicalmaxindex = 0;
for(i=0;i<numseq;i++)
{
m1 = ajSeqcvtGetCodeK(cvt, seqcharptr[i][k]);
if(identical[m1] > identical[identicalmaxindex])
identicalmaxindex = m1;
}
for(i=0;i<numseq;i++)
{
m1 = ajSeqcvtGetCodeK(cvt, seqcharptr[i][k]);
if(matching[m1] > matching[matchingmaxindex])
matchingmaxindex = m1;
else if(matching[m1] == matching[matchingmaxindex])
{
if(identical[m1] > identical[matchingmaxindex])
matchingmaxindex= m1;
}
}
iscons = ajFalse;
boxindex = -1;
max = -3;
ajDebug("k:%2d highindex:%2d matching:%4.2f\n",
k, highindex,
matching[ajSeqcvtGetCodeK(cvt, seqcharptr[highindex][k])]);
if(highindex != -1 &&
matching[ajSeqcvtGetCodeK(cvt, seqcharptr[highindex][k])] >= fplural)
{
iscons = ajTrue;
boxindex = highindex;
}
else
{
for(i=0;i<numseq;i++)
{
m1 = ajSeqcvtGetCodeK(cvt, seqcharptr[i][k]);
if(matching[m1] > max)
{
max = matching[m1];
highindex = i;
}
else if(matching[m1] == max)
{
if(identical[m1] >
identical[ajSeqcvtGetCodeK(cvt,
seqcharptr[highindex][k])] )
{
max = matching[m1];
highindex = i;
}
}
}
if(matching[ajSeqcvtGetCodeK(cvt,
seqcharptr[highindex][k])] >= fplural)
{
iscons = ajTrue;
boxindex = highindex;
}
}
if(iscons)
{
if(!collision)
{
/* check for collisions */
if(alternative == 1)
{
/* check to see if this is unique for collisions */
for(i=0;i<numseq;i++)
{
m1 = ajSeqcvtGetCodeK(cvt, seqcharptr[i][k]);
if(identical[m1] >= identical[identicalmaxindex] &&
m1 != identicalmaxindex)
iscons = ajFalse;
}
/*ajDebug("after (alt=1) iscons: %B",iscons);*/
}
else if(alternative == 2)
{
for(i=0;i<numseq;i++)
{
m1 = ajSeqcvtGetCodeK(cvt, seqcharptr[i][k]);
if((matching[m1] >= matching[matchingmaxindex] &&
m1 != matchingmaxindex &&
matrix[m1][matchingmaxindex] < 0.1)||
(identical[m1] >= identical[matchingmaxindex]
&& m1 != matchingmaxindex))
iscons = ajFalse;
}
}
else if(alternative == 3)
{
/*
** to do this check one is NOT in consensus to see if
** another score of fplural has been found
*/
ms = ajSeqcvtGetCodeK(cvt, seqcharptr[highindex][k]);
for(i=0;i<numseq;i++)
{
m1 = ajSeqcvtGetCodeK(cvt, seqcharptr[i][k]);
if(ms != m1 && colcheck[m1] == 0.0)
/* NOT in the current consensus */
for(j=0;j<numseq;j++)
{
m2 = ajSeqcvtGetCodeK(cvt, seqcharptr[j][k]);
if( matrix[ms][m2] < 0.1)
{
/* NOT in the current consensus */
if( matrix[m1][m2] > 0.1)
colcheck[m1] +=
ajSeqsetGetseqWeight(seqset,
j);
}
}
}
for(i=0;i<numseq;i++)
{
m1 = ajSeqcvtGetCodeK(cvt, seqcharptr[i][k]);
/* if any other matches then we have a collision */
if(colcheck[m1] >= fplural)
iscons = ajFalse;
}
/*ajDebug("after alt=2 iscons: %B", iscons);*/
}
else
{
for(i=0;i<numseq;i++)
{
m1 = ajSeqcvtGetCodeK(cvt, seqcharptr[i][k]);
if((matching[m1] >= matching[matchingmaxindex] &&
m1 != matchingmaxindex &&
matrix[m1][matchingmaxindex] < 0.1))
iscons = ajFalse;
if(identical[m1] >= identical[matchingmaxindex] &&
m1 != matchingmaxindex &&
matrix[m1][matchingmaxindex] > 0.1)
iscons = ajFalse;
}
if(!iscons)
{ /* matches failed try identicals */
if(identical[identicalmaxindex] >= fplural)
{
iscons = ajTrue;
/*
** if nothing has an equal or higher match that
** does not match highest then false
*/
for(i=0;i<numseq;i++)
{
m1 = ajSeqcvtGetCodeK(cvt, seqcharptr[i][k]);
if(identical[m1] >=
identical[identicalmaxindex] &&
m1 != identicalmaxindex)
iscons = ajFalse;
else if(matching[m1] >=
matching[identicalmaxindex] &&
matrix[m1][matchingmaxindex] <= 0.0)
iscons = ajFalse;
else if(m1 == identicalmaxindex)
j = i;
}
if(iscons)
highindex = j;
}
}
}
}
if(identity)
{
j = 0;
for(i=0;i<numseq;i++)
if(seqcharptr[highindex][k] == seqcharptr[i][k])
j++;
if(j<identity)
iscons = ajFalse;
}
}
/*
** Done a full line of residues
** Boxes have been defined up to this point
*/
if(count >= numres )
{
/* check y position for next set */
y=y-(yincr*((float)numseq+(float)2.0+((float)consensus*(float)2)));
if(y<yincr*((float)numseq+(float)2.0+((float)consensus*(float)2)))
{
/* full page - print it */
y=ystart-(float)6.0;
startseq = 0;
endseq = seqperpage;
newILstart = newILend;
newILend = k;
while(startseq < numseq)
{
/* AJB */
/*if(startseq != 0)
ajGraphNewpage(graph, AJFALSE);*/
/*ajDebug("Inner loop: startseq: %d numseq: %d endseq: %d\n",
startseq, numseq, endseq);*/
if(endseq>numseq)
endseq=numseq;
prettyplot_fillinboxes(numseq,ajSeqsetGetLen(seqset),
startseq,endseq,
newILstart,newILend,
numres,resbreak,
boxit,boxcol,consensus,
ystart,yincr,cvt);
startseq = endseq;
endseq += seqperpage;
ajGraphNewpage(graph, AJFALSE);
}
}
count = 0;
gapcount = 0;
}
count++;
countforgap++;
for(j=0;j<numseq;j++)
{
/* START OF BOXES */
if(boxit)
{
seqboxptr[j][k] = 0;
if(boxindex!=-1)
{
myindex = boxindex;
if(matrix[ajSeqcvtGetCodeK(cvt, seqcharptr[j][k])]
[ajSeqcvtGetCodeK(cvt, seqcharptr[myindex][k])] > 0)
part = 1.0;
else
{
if(identical[ajSeqcvtGetCodeK(cvt,
seqcharptr[j][k])] >=
fplural)
part = 1.0;
else
part = 0.0;
}
if(previous[j] != part)
/* draw vertical line */
seqboxptr[j][k] |= BOXLEF;
if(j==0)
{
/* special case for horizontal line */
if(part)
{
currentstate = 1;
/* draw hori line */
seqboxptr[j][k] |= BOXTOP;
}
else
currentstate = 0;
}
else
{
/* j != 0 Normal case for horizontal line */
if(part != currentstate)
{
/*draw hori line */
seqboxptr[j][k] |= BOXTOP;
currentstate = (ajint) part;
}
}
if(j== numseq-1 && currentstate)
/* draw horiline at bottom */
seqboxptr[j][k] |= BOXBOT;
previous[j] = (ajint) part;
}
else
{
part = 0;
if(previous[j])
{
/* draw vertical line */
seqboxptr[j][k] |= BOXLEF;
}
previous[j] = 0;
}
if(count == numres || k == kmax || countforgap >= resbreak )
{ /* last one on the row or a break*/
if(previous[j])
{
/* draw vertical line */
seqboxptr[j][k] |= BOXRIG;
}
previous[j] = 0;
}
} /* end box */
if(boxit && boxcol)
if(boxindex != -1)
{
myindex = boxindex;
if(matrix[ajSeqcvtGetCodeK(cvt, seqcharptr[j][k])]
[ajSeqcvtGetCodeK(cvt, seqcharptr[myindex][k])] > 0
|| identical[ajSeqcvtGetCodeK(cvt, seqcharptr[j][k])] >=
fplural )
seqboxptr[j][k] |= BOXCOLOURED;
}
/* END OF BOXES */
if(ajSeqcvtGetCodeK(cvt, seqcharptr[j][k]))
res[0] = seqcharptr[j][k];
else
res[0] = '-';
if(colourbyconsensus)
{
part = (float) matrix[ajSeqcvtGetCodeK(cvt, seqcharptr[j][k])]
[ajSeqcvtGetCodeK(cvt, seqcharptr[highindex][k])];
if(iscons && seqcharptr[highindex][k] == seqcharptr[j][k])
seqcolptr[j][k] = cidentity;
else if(part > 0.0)
seqcolptr[j][k] = csimilarity;
else
seqcolptr[j][k] = cother;
}
else if(colourbyresidues)
seqcolptr[j][k] = colmat[ajSeqcvtGetCodeK(cvt,
seqcharptr[j][k])];
else if(iscons && colourbyshade)
{
part = (float) matrix[ajSeqcvtGetCodeK(cvt, seqcharptr[j][k])]
[ajSeqcvtGetCodeK(cvt, seqcharptr[highindex][k])];
if(part >= 1.5)
seqcolptr[j][k] = shadecolour[0];
else if(part >= 1.0)
seqcolptr[j][k] = shadecolour[1];
else if(part >= 0.5)
seqcolptr[j][k] = shadecolour[2];
else
seqcolptr[j][k] = shadecolour[3];
}
else if(colourbyshade)
seqcolptr[j][k] = shadecolour[3];
else
seqcolptr[j][k] = BLACK;
}
if(consensus)
{
if(iscons)
res[0] = seqcharptr[highindex][k];
else
res[0] = '-';
strcat(constr,res);
}
if(countforgap >= resbreak)
{
gapcount++;
countforgap=0;
}
}
startseq = 0;
endseq=seqperpage;
newILstart = newILend;
newILend = k;
while(startseq < numseq)
{
if(startseq)
ajGraphNewpage(graph, AJFALSE);
/*ajDebug("Final loop: startseq: %d numseq: %d endseq: %d\n",
startseq, numseq, endseq);*/
if(endseq>numseq)
endseq = numseq;
prettyplot_fillinboxes(numseq,ajSeqsetGetLen(seqset),
startseq,endseq,
newILstart,newILend,
numres,resbreak,
boxit,boxcol,consensus,
ystart,yincr,cvt);
startseq = endseq;
endseq += seqperpage;
}
ajGraphicsGetCharsize(&defheight,¤tscale);
if(boxit && boxcol)
oldfg = ajGraphicsSetFgcolour(oldfg);
ajGraphicsCloseWin();
ajGraphxyDel(&graph);
ajStrDel(&sidentity);
ajStrDel(&ssimilarity);
ajStrDel(&sother);
ajStrDel(&options);
ajStrDel(&altstr);
ajStrDel(&matcodes);
for(i=0;i<numseq;i++)
{
ajStrDel(&seqnames[i]);
AJFREE(seqcolptr[i]);
if(seqboxptr)
AJFREE(seqboxptr[i]);
}
AJFREE(seqcolptr);
AJFREE(seqboxptr);
AJFREE(seqnames);
AJFREE(score);
AJFREE(previous);
AJFREE(seqcount);
AJFREE(colmat);
AJFREE(shadecolour);
freeptr = (void *) seqcharptr;
AJFREE(freeptr);
AJFREE(identical);
AJFREE(matching);
AJFREE(colcheck);
ajSeqsetDel(&seqset);
ajMatrixDel(&cmpmatrix);
ajStrDel(&shade);
ajStrDel(&sboxcolval);
ajStrDel(&sidentity);
ajStrDel(&ssimilarity);
ajStrDel(&sother);
ajFloatDel(&pair);
ajTimeDel(&ajtime);
AJFREE(constr);
embExit();
return 0;
}
int main(int argc, char **argv)
{
AjPSeqset seqset;
AjPSeqall seqall;
AjPSeq queryseq;
const AjPSeq targetseq;
ajint wordlen;
AjPTable wordsTable = NULL;
AjPList* matchlist = NULL;
AjPFile logfile;
AjPFeattable* seqsetftables = NULL;
AjPFeattable seqallseqftable = NULL;
AjPFeattabOut ftoutforseqsetseq = NULL;
AjPFeattabOut ftoutforseqallseq = NULL;
AjPAlign align = NULL;
AjIList iter = NULL;
ajint targetstart;
ajint querystart;
ajint len;
ajuint i, j;
ajulong nAllMatches = 0;
ajulong sumAllScore = 0;
AjBool dumpAlign = ajTrue;
AjBool dumpFeature = ajTrue;
AjBool checkmode = ajFalse;
EmbPWordRK* wordsw = NULL;
ajuint npatterns = 0;
ajuint seqsetsize;
ajuint nmatches;
ajuint* nmatchesseqset;
ajuint* lastlocation; /* Cursors for Rabin-Karp search. */
/* Shows until what point the query sequence was
* scanned for a pattern sequences in the seqset.
*/
char* paddedheader = NULL;
const char* header;
AjPStr padding;
header = "Pattern %S #pat-sequences #all-matches avg-match-length\n";
padding = ajStrNew();
embInit("wordmatch", argc, argv);
wordlen = ajAcdGetInt("wordsize");
seqset = ajAcdGetSeqset("asequence");
seqall = ajAcdGetSeqall("bsequence");
logfile = ajAcdGetOutfile("logfile");
dumpAlign = ajAcdGetToggle("dumpalign");
dumpFeature = ajAcdGetToggle("dumpfeat");
if(dumpAlign)
{
align = ajAcdGetAlign("outfile");
ajAlignSetExternal(align, ajTrue);
}
seqsetsize = ajSeqsetGetSize(seqset);
ajSeqsetTrim(seqset);
AJCNEW0(matchlist, seqsetsize);
AJCNEW0(seqsetftables, seqsetsize);
AJCNEW0(nmatchesseqset, seqsetsize);
if (dumpFeature)
{
ftoutforseqsetseq = ajAcdGetFeatout("aoutfeat");
ftoutforseqallseq = ajAcdGetFeatout("boutfeat");
}
checkmode = !dumpFeature && !dumpAlign;
embWordLength(wordlen);
ajFmtPrintF(logfile, "Small sequence/file for constructing"
" target patterns: %S\n", ajSeqsetGetUsa(seqset));
ajFmtPrintF(logfile, "Large sequence/file to be scanned"
" for patterns: %S\n", ajSeqallGetUsa(seqall));
ajFmtPrintF(logfile, "Number of sequences in the patterns file: %u\n",
seqsetsize);
ajFmtPrintF(logfile, "Pattern/word length: %u\n", wordlen);
for(i=0;i<seqsetsize;i++)
{
targetseq = ajSeqsetGetseqSeq(seqset, i);
embWordGetTable(&wordsTable, targetseq);
}
AJCNEW0(lastlocation, seqsetsize);
if(ajTableGetLength(wordsTable)>0)
{
npatterns = embWordRabinKarpInit(wordsTable,
&wordsw, wordlen, seqset);
ajFmtPrintF(logfile, "Number of patterns/words found: %u\n", npatterns);
while(ajSeqallNext(seqall,&queryseq))
{
for(i=0;i<seqsetsize;i++)
{
lastlocation[i]=0;
if (!checkmode)
matchlist[i] = ajListstrNew();
}
nmatches = embWordRabinKarpSearch(
ajSeqGetSeqS(queryseq), seqset,
(EmbPWordRK const *)wordsw, wordlen, npatterns,
matchlist, lastlocation, checkmode);
nAllMatches += nmatches;
if (checkmode)
continue;
for(i=0;i<seqsetsize;i++)
{
if(ajListGetLength(matchlist[i])>0)
{
iter = ajListIterNewread(matchlist[i]) ;
while(embWordMatchIter(iter, &targetstart, &querystart, &len,
&targetseq))
{
if(dumpAlign)
{
ajAlignDefineSS(align, targetseq, queryseq);
ajAlignSetScoreI(align, len);
/* ungapped alignment means same length
* for both sequences
*/
ajAlignSetSubRange(align, targetstart, 1, len,
ajSeqIsReversed(targetseq),
ajSeqGetLen(targetseq),
querystart, 1, len,
ajSeqIsReversed(queryseq),
ajSeqGetLen(queryseq));
}
}
if(dumpAlign)
{
ajAlignWrite(align);
ajAlignReset(align);
}
if(ajListGetLength(matchlist[i])>0 && dumpFeature)
{
embWordMatchListConvToFeat(matchlist[i],
&seqsetftables[i],
&seqallseqftable,
targetseq, queryseq);
ajFeattableWrite(ftoutforseqallseq, seqallseqftable);
ajFeattableDel(&seqallseqftable);
}
ajListIterDel(&iter);
}
embWordMatchListDelete(&matchlist[i]);
}
}
/* search completed, now report statistics */
for(i=0;i<npatterns;i++)
{
sumAllScore += wordsw[i]->lenMatches;
for(j=0;j<wordsw[i]->nseqs;j++)
nmatchesseqset[wordsw[i]->seqindxs[j]] +=
wordsw[i]->nSeqMatches[j];
}
ajFmtPrintF(logfile, "Number of sequences in the file scanned "
"for patterns: %u\n", ajSeqallGetCount(seqall));
ajFmtPrintF(logfile, "Number of all matches: %Lu"
" (wordmatch finds exact matches only)\n", nAllMatches);
if(nAllMatches>0)
{
ajFmtPrintF(logfile, "Sum of match lengths: %Lu\n", sumAllScore);
ajFmtPrintF(logfile, "Average match length: %.2f\n",
sumAllScore*1.0/nAllMatches);
ajFmtPrintF(logfile, "\nDistribution of the matches among pattern"
" sequences:\n");
ajFmtPrintF(logfile, "-----------------------------------------"
"-----------\n");
for(i=0;i<ajSeqsetGetSize(seqset);i++)
{
if (nmatchesseqset[i]>0)
ajFmtPrintF(logfile, "%-42s: %8u\n",
ajSeqGetNameC(ajSeqsetGetseqSeq(seqset, i)),
nmatchesseqset[i]);
ajFeattableWrite(ftoutforseqsetseq, seqsetftables[i]);
ajFeattableDel(&seqsetftables[i]);
}
ajFmtPrintF(logfile, "\nPattern statistics:\n");
ajFmtPrintF(logfile, "-------------------\n");
if(wordlen>7)
ajStrAppendCountK(&padding, ' ', wordlen-7);
paddedheader = ajFmtString(header,padding);
ajFmtPrintF(logfile, paddedheader);
for(i=0;i<npatterns;i++)
if (wordsw[i]->nMatches>0)
ajFmtPrintF(logfile, "%-7s: %12u %12u %17.2f\n",
wordsw[i]->word->fword, wordsw[i]->nseqs,
wordsw[i]->nMatches,
wordsw[i]->lenMatches*1.0/wordsw[i]->nMatches);
}
}
for(i=0;i<npatterns;i++)
{
for(j=0;j<wordsw[i]->nseqs;j++)
AJFREE(wordsw[i]->locs[j]);
AJFREE(wordsw[i]->locs);
AJFREE(wordsw[i]->seqindxs);
AJFREE(wordsw[i]->nnseqlocs);
AJFREE(wordsw[i]->nSeqMatches);
AJFREE(wordsw[i]);
}
embWordFreeTable(&wordsTable);
AJFREE(wordsw);
AJFREE(matchlist);
AJFREE(lastlocation);
AJFREE(nmatchesseqset);
AJFREE(seqsetftables);
if(dumpAlign)
{
ajAlignClose(align);
ajAlignDel(&align);
}
if(dumpFeature)
{
ajFeattabOutDel(&ftoutforseqsetseq);
ajFeattabOutDel(&ftoutforseqallseq);
}
ajFileClose(&logfile);
ajSeqallDel(&seqall);
ajSeqsetDel(&seqset);
ajSeqDel(&queryseq);
ajStrDel(&padding);
AJFREE(paddedheader);
embExit();
return 0;
}
