@Test
public void testToString() {
assertTrue(
"(ipv4) toString should return expected format",
domain1.toString().equals("200.193.193.in-addr.arpa(193.193.200.0/24 INADDR not-dashed)"));
assertTrue(
"(ipv6) toString should return expected format",
domain4.toString().equals("2.1.2.1.5.5.5.2.0.2.1.e164.arpa(E164 not-dashed)"));
}
// <seq id> <alignment start> <alignment end> <envelope start> <envelope end>
// <hmm acc> <hmm name> <type> <hmm start> <hmm end> <hmm length> <bit score> <E-value>
// <significance> <clan>
public void pfamscan2xdom() {
TreeMap<Integer, Domain> currentDoms = new TreeMap<Integer, Domain>();
Pattern comment = Pattern.compile("^#.*");
Pattern empty = Pattern.compile("^$");
try {
FileWriter fw = new FileWriter(outfile);
String line;
FileInputStream fis = new FileInputStream(this.domtblout);
DataInputStream dis = new DataInputStream(fis);
BufferedReader br = new BufferedReader(new InputStreamReader(dis));
String currentId = null;
StringBuilder xdom = new StringBuilder();
int didField = 6;
if (accMode) didField = 5;
while ((line = br.readLine()) != null) {
if (comment.matcher(line).matches()) continue;
if (empty.matcher(line).matches()) continue;
String[] fields = line.split("\\s+");
// 0 -> protein id
// 5 -> domain acc
// 6 -> domain name (id)
// 12 -> domain evalue
// 8, 9 -> hmm coord
// 1, 2 -> align coord
// 3, 4 -> env coord
// 15 -> clanid
String thisId = fields[0];
Pattern p = Pattern.compile("\\w+.\\d+");
Matcher m = p.matcher(thisId);
if (m.find()) {
String[] pidFields = thisId.split("\\.");
thisId = pidFields[0];
}
if ((currentId != null) && (!thisId.equals(currentId))) {
if (xdom.length() != 0) {
fw.write(xdom.toString() + "\n");
// merge split hits
if (merge) currentDoms = mergeHits(currentDoms);
// resolve overlaps
if (resolveOverlaps) resolveOverlaps(currentDoms, null);
if (collapse) currentDoms = collapseRepeats(currentDoms);
// write the rest of the domains
for (int key : currentDoms.keySet()) {
Domain cdom = currentDoms.get(key);
if ((cdom = currentDoms.get(key)) != null) fw.write(cdom.toString() + "\n");
}
}
xdom.setLength(0);
currentDoms.clear();
}
if (xdom.length() == 0) {
currentId = fields[0];
p = Pattern.compile("\\w+.\\d+");
m = p.matcher(currentId);
if (m.find()) {
String[] pidFields = currentId.split("\\.");
currentId = pidFields[0];
}
xdom.setLength(0);
xdom.append(">" + currentId);
}
if (evalue != null) if (Double.parseDouble(fields[12]) > evalue) continue;
// ensure that the version number is removed if we are
// in acc mode
String did = fields[didField];
if (accMode) {
p = Pattern.compile("PF\\d+.\\d+");
m = p.matcher(did);
if (m.find()) {
String[] didFields = did.split("\\.");
did = didFields[0];
}
}
if (clanMode) {
p = Pattern.compile("CL\\d+");
m = p.matcher(fields[14]);
if (m.find()) did = fields[14];
}
Domain dom =
new Domain(
did,
Integer.parseInt(fields[1]),
Integer.parseInt(fields[2]),
Integer.parseInt(fields[8]),
Integer.parseInt(fields[9]),
Double.parseDouble(fields[12]));
currentDoms.put(Integer.parseInt(fields[1]), dom);
}
if (xdom.length() != 0) {
fw.write(xdom.toString() + "\n");
if (merge) currentDoms = mergeHits(currentDoms);
if (resolveOverlaps) resolveOverlaps(currentDoms, null);
if (collapse) currentDoms = collapseRepeats(currentDoms);
for (int key : currentDoms.keySet()) {
Domain cdom = currentDoms.get(key);
if ((cdom = currentDoms.get(key)) != null) fw.write(cdom.toString() + "\n");
}
}
fis.close();
dis.close();
br.close();
fw.close();
} catch (Exception e) {
e.printStackTrace();
}
}