public BlastResult performTBlastn(File pssm, File db) {
Runtime rt = Runtime.getRuntime();
BlastResult result = null;
try {
String db_path = db.getAbsolutePath();
// db_path = db_path.substring(0, db_path.lastIndexOf('.')-1);
String commandline =
tblastn_exec_path
+ " -in_pssm "
+ pssm.getAbsolutePath()
+ " -db "
+ db_path
+ " -outfmt 6";
Process pr = rt.exec(commandline);
BufferedReader in = new BufferedReader(new InputStreamReader(pr.getInputStream()));
String line = null;
double minEvalue = 0.01;
while ((line = in.readLine()) != null) {
BlastResult r = new BlastResult();
r.parse(line);
if (r.getEvalue() < minEvalue) {
result = r;
minEvalue = r.getEvalue();
}
}
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
return result;
}
/**
* @param args
* @throws IOException
*/
public static void main(String[] args) throws IOException {
MassiveTblastnPicorna mbxp = new MassiveTblastnPicorna();
List<String> subjectGI =
mbxp.getSubjectGI(
new File("/home/javier/Dropbox/Posdoc/PICORNAVIRIDAE/separateByProt/queriesGI"));
Map<String, String> genomes =
mbxp.loadGenomes(
subjectGI, new File("/home/javier/Dropbox/Posdoc/PICORNAVIRIDAE/PICORNAVIRIDAE.fas"));
File tmp_fasta = new File("/home/javier/Dropbox/Posdoc/PICORNAVIRIDAE/separateByProt/tmp.fas");
File tmp_db = new File("/home/javier/Dropbox/Posdoc/PICORNAVIRIDAE/separateByProt/tmp");
PrintStream out =
new PrintStream(
"/home/javier/Dropbox/Posdoc/PICORNAVIRIDAE/separateByProt/tblastn.results");
List<File> pssms = new ArrayList<File>();
for (int i = 1; i < 12; i++) {
String number = String.format("%02d", i);
pssms.add(
new File(
"/home/javier/Dropbox/Posdoc/PICORNAVIRIDAE/separateByProt/ORF" + number + ".pssm"));
}
for (String string : subjectGI) {
// Iterate over each genome
String genome_sequence = genomes.get(string);
mbxp.exportTmpFasta(genome_sequence, string, tmp_fasta);
mbxp.makeblastdb(tmp_fasta, tmp_db);
for (int i = 0; i < 11; i++) {
// Iterate over each protein
File orf_pssm = pssms.get(i);
BlastResult result = mbxp.performTBlastn(orf_pssm, tmp_db);
if (result != null) {
String out_seq;
int beginIndex = Math.min(result.getSubjectStart(), result.getSubjectEnd());
int endIndex = Math.max(result.getSubjectStart(), result.getSubjectEnd());
out_seq = genome_sequence.substring(beginIndex - 1, endIndex);
if (result.getSubjectStart() > result.getSubjectEnd()) {
out_seq = Complementary.reverseComplementary(out_seq);
}
out.println(">" + string + "|ORF" + String.format("%02d", i + 1));
out.println(out_seq);
}
}
System.out.println("Genome " + string + " Done.");
}
out.close();
}