//$MENU window=main;popup=Sequence;insertpopupafter=Scripts;Name=Translate selected sequences

integer i,j,k,x,tl,ct,frame,nuclseqlen,ok,baseidx,vl;
string seqlist,expername,translexpername,transtablelist,transtable,st,st2,st3;
string key,nuclseq,aminoseq,base,transcode,kodon,amino;
DIALOG dlg;
FILE fp;
XMLNODE xmldoc,xmlnd;
SEQUENCE sq,tsq;
INDEX kodonidx;

for i=1 to DbGetExperCount do
   if DbGetExperClass(i)="SEQ" then {
      if find(DbGetExperName(i),"_TRANSL",1)=0 then
         seqlist=seqlist+DbGetExperName(i)+"	";
   }
expername=seqlist;expername=splitstring(expername,"	");
DlgAddText(dlg,"Select the experiment you want to translate",15,15,230,15);
DlgAddList(dlg,seqlist,expername,15,35,150,100,"LIST");

if not(FileOpenRead(fp,DbGetStartupDir+"\aminotrans.txt")) then {
   Message("File 'aminotrans.txt' will now be copied");
   ExtractAttach("trans",DbGetStartupDir+"\aminotrans.txt");
   if not(FileOpenRead(fp,DbGetStartupDir+"\aminotrans.txt")) then {
      message("Could not find the file 'aminotrans.txt'");
      stop;
   }
}
while not(fileIsEnd(fp)) do {
   st=FileRead(fp,999);
   if length(st)>5 then transtablelist=transtablelist+splitstring(st,"	")+"	";
}
FileClose(fp);
transtable="Standard";
DlgAddText(dlg,"Select the translation table",15,150,150,15);
DlgAddList(dlg,transtablelist,transtable,15,170,250,150,"LIST");


x=DlgShow(dlg,"Translate selected sequences",380,350);
if x<>1 then stop;

//create translated experiment if necessary
translexpername=expername+"_TRANSL";
if DbGetExperNr(translexpername)=0 then {
   DbCreateExperType(translexpername,"SEQ");
   XMLNodeCreateDoc("Settings",xmldoc,xmlnd);
   XMLNodeAddText(XMLNodeAddChild(xmlnd,"Nucl"),"0");
   DbSetExperSettings(DbGetExperNr(translexpername),xmldoc);
   XMLNodeCloseDocument(xmldoc);
}

//create translation frame field if necessary
if DbGetFieldNr("TranslationFrames")=0 then DbAddField("TranslationFrames");

//load translation info
FileOpenRead(fp,DbGetStartupDir+"\aminotrans.txt");
transcode="";
while not(fileIsEnd(fp)) do {
   st=FileRead(fp,999);
   if splitstring(st,"	")=transtable then transcode=st;
}
FileClose(fp);
if transcode="" then {
   message("Unable to load the translation info");
   stop;
}

//build kodon index;
for i=0 to 3 do {
   for j=0 to 3 do {
      for k=0 to 3 do {
         kodon="";
         if i=0 then kodon=kodon+"T";
         if i=1 then kodon=kodon+"C";
         if i=2 then kodon=kodon+"A";
         if i=3 then kodon=kodon+"G";
         if j=0 then kodon=kodon+"T";
         if j=1 then kodon=kodon+"C";
         if j=2 then kodon=kodon+"A";
         if j=3 then kodon=kodon+"G";
         if k=0 then kodon=kodon+"T";
         if k=1 then kodon=kodon+"C";
         if k=2 then kodon=kodon+"A";
         if k=3 then kodon=kodon+"G";
         vl=(i*4+j)*4+k+1;
         IdxSet(kodonidx,kodon,vl);
      }
   }
}

setbusy("Performing translations...");
for tl=1 to DbGetEntryCount do if DbGetSel(DbGetEntryKey(tl)) then {
   ct=ct+1;
   if ct/10=floor(ct/10) then setbusy("Performing translations... - "+str(ct,0,0));
   key=DbGetEntryKey(tl);
   frame=1;
   st=DbGetField(key,"TranslationFrames");
   while st<>"" do {
      st2=splitstring(st,":=");
      st3=splitstring(st,"|");
      if st2=expername then frame=val(st3);
   }
   if frame<1 then frame=1; if frame>3 then frame=3;
   if SeqLoad(sq,key,expername) then {
      SeqGet(sq,nuclseq);nuclseqlen=length(nuclseq);
      aminoseq="";
      i=frame;
      while i+3<=nuclseqlen do {
         baseidx=0;
         kodon=substring(nuclseq,i,i+2);
         baseidx=IdxGet(kodonidx,kodon);
         if baseidx>0 then amino=substring(transcode,baseidx,baseidx);
         else amino="X";
         if amino="*" then amino="X";
         aminoseq=aminoseq+amino;
         i=i+3;
      }
      if not(SeqLoad(tsq,key,translexpername)) then
         SeqCreate(tsq,translexpername,"",key);
      SeqSet(tsq,aminoseq);
      SeqSave(tsq);
   }
}

setbusy("");
#_#_#_#_#_#_# ATTACHMENTS #_#_#_#_#_#_#_
>>ATTACHMENT V_1.0 trans 69
FDHEGBGOGEGBHCGEAJEGEGEMEMFDFDFDFDFJFJCK
CKEDEDCKFHEMEMEMEMFAFAFAFAEIEIFBFBFCFCFC
FCEJEJEJENFEFEFEFEEOEOELELFDFDFCFCFGFGFG
FGEBEBEBEBEEEEEFEFEHEHEHEHANAKFGGFHCHEGF
GCHCGBHEGFCAENGJHEGPGDGIGPGOGEHCGJGBGMAJ
EGEGEMEMFDFDFDFDFJFJCKCKEDEDFHFHEMEMEMEM
FAFAFAFAEIEIFBFBFCFCFCFCEJEJENENFEFEFEFE
EOEOELELFDFDCKCKFGFGFGFGEBEBEBEBEEEEEFEF
EHEHEHEHANAKFJGFGBHDHECAENGJHEGPGDGIGPGO
GEHCGJGBGMAJEGEGEMEMFDFDFDFDFJFJCKCKEDED
FHFHFEFEFEFEFAFAFAFAEIEIFBFBFCFCFCFCEJEJ
ENENFEFEFEFEEOEOELELFDFDFCFCFGFGFGFGEBEB
EBEBEEEEEFEFEHEHEHEHANAKENGPGMGECAENGJHE
GPGDGIGPGOGEHCGJGBGMCAGBGOGECAENHJGDGPHA
GMGBHDGNGBAJEGEGEMEMFDFDFDFDFJFJCKCKEDED
FHFHEMEMEMEMFAFAFAFAEIEIFBFBFCFCFCFCEJEJ
EJENFEFEFEFEEOEOELELFDFDFCFCFGFGFGFGEBEB
EBEBEEEEEFEFEHEHEHEHANAKEJGOHGGFHCHEGFGC
HCGBHEGFCAENGJHEGPGDGIGPGOGEHCGJGBGMAJEG
EGEMEMFDFDFDFDFJFJCKCKEDEDFHFHEMEMEMEMFA
FAFAFAEIEIFBFBFCFCFCFCEJEJENENFEFEFEFEEO
EOELELFDFDFDFDFGFGFGFGEBEBEBEBEEEEEFEFEH
EHEHEHANAKEDGJGMGJGBHEGFCAENGBGDHCGPGOHF
GDGMGFGBHCCAGBGOGECAEEGBHJGDGMGBGEGBGDGF
GBGOAJEGEGEMEMFDFDFDFDFJFJFBFBEDEDCKFHEM
EMEMEMFAFAFAFAEIEIFBFBFCFCFCFCEJEJEJENFE
FEFEFEEOEOELELFDFDFCFCFGFGFGFGEBEBEBEBEE
EEEFEFEHEHEHEHANAKFAHCGPHEGPHKGPGBGOCAEN
GJHEGPGDGIGPGOGEHCGJGBGMCACIGBGOGECAELGJ
GOGFHEGPHAGMGBHDHECJAJEGEGEMEMFDFDFDFDFJ
FJCKCKEDEDFHFHEMEMEMEMFAFAFAFAEIEIFBFBFC
FCFCFCEJEJEJENFEFEFEFEEOEOELELFDFDFCFCFG
FGFGFGEBEBEBEBEEEEEFEFEHEHEHEHANAKFAGMGB
GOHECAENGJHEGPGDGIGPGOGEHCGJGBGMCPEDGIGM
GPHCGPHAGMGBHDHEAJEGEGEMEMFDFDFDFDFJFJCK
CKEDEDCKFHEMEMEMEMFAFAFAFAEIEIFBFBFCFCFC
FHEJEJEJENFEFEFEFEEOEOELELFDFDFCFCFGFGFG
FGEBEBEBEBEEEEEFEFEHEHEHEHANAKEFGDGIGJGO
GPGEGFHCGNCAENGJHEGPGDGIGPGOGEHCGJGBGMAJ
EGEGEMEMFDFDFDFDFJFJCKCKEDEDFHFHEMEMEMEM
FAFAFAFAEIEIFBFBFCFCFCFCEJEJEJENFEFEFEFE
EOEOEOELFDFDFDFDFGFGFGFGEBEBEBEBEEEEEFEF
EHEHEHEHANAKEFHFHAGMGPHEGJGECAENGBGDHCGP
GOHFGDGMGFGBHCAJEGEGEMEMFDFDFDFDFJFJCKFB
EDEDEDFHEMEMEMEMFAFAFAFAEIEIFBFBFCFCFCFC
EJEJEJENFEFEFEFEEOEOELELFDFDFCFCFGFGFGFG
EBEBEBEBEEEEEFEFEHEHEHEHANAKEFHFGCGBGDHE
GFHCGJGBGMAJEGEGEMEMFDFDFDFDFJFJCKCKEDED
CKFHEMEMEMEMFAFAFAFAEIEIFBFBFCFCFCFCEJEJ
EJENFEFEFEFEEOEOELELFDFDFCFCFGFGFGFGEBEB
EBEBEEEEEFEFEHEHEHEHANAKEHHCGPHFHACAEJEJ
CAHJGFGBHDHEHDAJEGEGEMEMFDFDFDFDFJFJCKCK
EDEDCKFHEMEMEMFDFAFAFAFAEIEIFBFBFCFCFCFC
EJEJEJENFEFEFEFEEOEOELELFDFDFCFCFGFGFGFG
EBEBEBEBEEEEEFEFEHEHEHEHANAKEBHDGDGJGEGJ
GBGOCAENGJHEGPGDGIGPGOGEHCGJGBGMAJEGEGEM
EMFDFDFDFDFJFJCKCKEDEDFHFHEMEMEMEMFAFAFA
FAEIEIFBFBFCFCFCFCEJEJENENFEFEFEFEEOEOEL
ELFDFDEHEHFGFGFGFGEBEBEBEBEEEEEFEFEHEHEH
EHANAKEGGMGBHEHHGPHCGNCAENGJHEGPGDGIGPGO
GEHCGJGBGMAJEGEGEMEMFDFDFDFDFJFJFJCKEDED
FHFHEMEMEMEMFAFAFAFAEIEIFBFBFCFCFCFCEJEJ
EJENFEFEFEFEEOEOEOELFDFDFDFDFGFGFGFGEBEB
EBEBEEEEEFEFEHEHEHEHANAKECGMGFHAGIGBHCGJ
HDGNGBCAENGBGDHCGPGOHFGDGMGFGBHCAJEGEGEM
EMFDFDFDFDFJFJCKFBEDEDCKFHEMEMEMEMFAFAFA
FAEIEIFBFBFCFCFCFCEJEJEJENFEFEFEFEEOEOEL
ELFDFDFCFCFGFGFGFGEBEBEBEBEEEEEFEFEHEHEH
EHANAK