Associated Region to Gene List

Script

import org.dom4j.Document;
import org.dom4j.Node;
import org.dom4j.Element;
import org.dom4j.io.SAXReader;

String getParameter(Element root, String parameterName) {
    List nodeList = root.selectNodes("Parameter[@name='" + parameterName + "']");
    if (nodeList.size() == 1) {
        Node parameterNode = (Node) nodeList.get(0);
        String parameter = "";
        if (parameterNode.isTextOnly())
            parameter = parameterNode.getText();
        else {
            List paramElements = parameterNode.elements();
            for (int i = 0; i < paramElements.size(); i++) {
                parameter += ((Element) paramElements.get(i)).asXML();
            }
        }
        if (parameter == null || parameter.equals(""))
            return null;
        else
            return parameter;
    } else if (nodeList.size() > 1) {
        throw new Exception("Multiple parameters matching " + parameterName + "!");
    }
    return null;
}

SAXReader reader = new SAXReader(false);
reader.setIncludeInternalDTDDeclarations(false);
reader.setIncludeExternalDTDDeclarations(false);

Document document = reader.read(new StringReader(paramXML));
Element root = document.getRootElement();

// Set parameter defaults.
nameType       = "associatedName";
addPvalue      = "false";
geneStatus     = "";
excludeHLA     = "false";
addRsId        = "false";
associatedSNPs = new ArrayList();
errorsOut      = errorsIn;

// Get parameters from XML.
String paramTmp;
// nameType.
if ((paramTmp = getParameter(root, "nameType")) != null) {
	if (paramTmp.equals("ensgId") || paramTmp.equals("ucscId") || paramTmp.equals("associatedName") || paramTmp.equals("description"))
		nameType = paramTmp;
	else {
		errorsOut.add("Name type '" + paramTmp + "' not recognised! Must be one of ensgId, ucscId, associatedName or description.");
		nameType = null;
	}
}
// addPvalue.
if ((paramTmp = getParameter(root, "addPvalue")) != null)
	addPvalue = paramTmp.equals("false") ? "false" : "true";
// Gene status.
if ((paramTmp = getParameter(root, "geneStatus")) != null) {
	geneStatus = paramTmp;
}
// excludeHLA.
if ((paramTmp = getParameter(root, "excludeHLA")) != null)
	excludeHLA = paramTmp.equals("false") ? "false" : "true";
// addRsId.
if ((paramTmp = getParameter(root, "addRsId")) != null)
	addRsId = paramTmp.equals("false") ? "false" : "true";
// associatedSNPs.
if ((paramTmp = getParameter(root, "associatedSNPs")) != null) {
    Document assocSnpDoc = reader.read(new StringReader(paramTmp));
    List nodeList = assocSnpDoc.selectNodes(xpath);
    for (int i = 0; i < nodeList.size(); i++) {
        Node node = (Node) nodeList.get(i);
        String xmlVal = node.asXML();
        if (xmlVal != null && ! xmlVal.equals(""))
            associatedSNPs.add(xmlVal);
    }
}

Value

Script

import org.dom4j.Document;
import org.dom4j.Node;
import org.dom4j.io.SAXReader;

SAXReader reader = new SAXReader(false);
reader.setIncludeInternalDTDDeclarations(false);
reader.setIncludeExternalDTDDeclarations(false);

List geneNames = new ArrayList();
for (int i = 0; i < associatedSNPs.size(); i++) {
	String xmlText = associatedSNPs.get(i);
	Document document = reader.read(new StringReader(xmlText));
	Node node = document.selectSingleNode("AssociatedSNP");
	if (node == null)
		continue;


	String line = addRsId.equals("true") ? node.attributeValue("rsId") + "\t" : "";
	// Get associated regions.
	List assocRegionNodes = node.selectNodes("AssociatedRegion");
	for (int j = 0; j < assocRegionNodes.size(); j++) {
		Node assocRegionNode = (Node) assocRegionNodes.get(j);

		if (assocRegionNode == null)
			continue;

		// Get associated Genes.
		Node assocGeneNode = assocRegionNode.selectSingleNode("AssociatedRegionGenes");
		if (assocGeneNode == null)
			continue;
		List geneNodes = assocGeneNode.selectNodes("Gene");
		for (int k = 0; k < geneNodes.size(); k++) {
			Node geneNode = (Node) geneNodes.get(k);
			if (geneNode == null)
				continue;
			String statusAtt = geneNode.attributeValue("status");
			if (! geneStatus.equals("") && ! statusAtt.equals(geneStatus))
				continue;

			Node locationNode = (Node) geneNode.selectSingleNode("Location[@ucscBuild='hg19']");
			if (locationNode != null && excludeHLA.equals("true")) {
				String chr = locationNode.attributeValue("chromosome");
				int start = locationNode.attributeValue("start") == null ? 0 : Integer.parseInt(locationNode.attributeValue("start"));
				int end   = locationNode.attributeValue("end")   == null ? 0 : Integer.parseInt(locationNode.attributeValue("end"));
				if (chr.equals("chr6") && start <= 33448354 && end >= 28477797)
					continue;
			}

			if (nameType.equals("ensgId")) {
				String id = line + geneNode.attributeValue("ensgId");
				if (addPvalue.equals("true"))
					id += node.attributeValue("pValue") == null ? "\t" : "\t" + node.attributeValue("pValue");
				geneNames.add(id);
				continue;
			}
			List geneNameNodes = geneNode.selectNodes("GeneName");
			for (int l = 0; l < geneNameNodes.size(); l++) {
				Node geneNameNode = (Node) geneNameNodes.get(l);
				if (geneNameNode == null)
					continue;
				String type = geneNameNode.attributeValue("type");
				if (type.equals(nameType)) {
					String name = line + geneNameNode.attributeValue("name");
					if (addPvalue.equals("true"))
						name += node.attributeValue("pValue") == null ? "\t" : "\t" + node.attributeValue("pValue");
					geneNames.add(name);
					continue;
				}
			}
		}
	}
}

Script

import org.dom4j.Document;
import org.dom4j.Element;
import org.dom4j.Node;
import org.dom4j.dom.DOMDocument;
import org.dom4j.io.SAXReader;

SAXReader reader = new SAXReader(false);
reader.setIncludeInternalDTDDeclarations(false);
reader.setIncludeExternalDTDDeclarations(false);

Document document = reader.read(new StringReader(associatedSNPs));
Element root = document.getRootElement();

// Create Document.
Document paramDoc = new DOMDocument();

// Add root element.
Element rootElement = paramDoc.addElement("Parameters");
Element paramElement = rootElement.addElement("Parameter")
	.addAttribute("name", "associatedSNPs");
paramElement.add(root.detach());

paramXML = paramDoc.asXML();

Script

List split = new ArrayList();
if (!string.equals("")) {
	String regexString = ",";
	if (regex != void) {
		regexString = regex;
	}
	String[] result = string.split(regexString);
	for (int i = 0; i < result.length; i++) {
		split.add(result[i]);
	}
}

Script

import org.dom4j.Element;
import org.dom4j.Document;
import org.dom4j.io.SAXReader;

List errors = new ArrayList();

if (additionalParams == void || additionalParams.size() == 0) {
	paramXmlOut = paramXmlIn;
} else {

	// Initialise Reader for parameter XML.
	SAXReader reader = new SAXReader(false);
	reader.setIncludeInternalDTDDeclarations(false);
	reader.setIncludeExternalDTDDeclarations(false);

	Document paramDoc = reader.read(new StringReader(paramXmlIn));
	Element rootElement = paramDoc.getRootElement();
	for (int i = 0; i < additionalParams.size(); i++) {
		String paramInfo = (String) additionalParams.get(i);
		if (paramInfo == null || paramInfo.equals(""))
			continue;
		String [] paramArray = paramInfo.split("\t");
		if (paramArray.length != 2) {
			errors.add("Invalid parameter '" + paramInfo + "' on line " + (i + 1) + ".");
			continue;
		}
		// Add Parameter.
		Element paramElement = rootElement.addElement("Parameter")
			.addAttribute("name", paramArray[0]);
		paramElement.addText(paramArray[1]);
	}
	paramXmlOut = paramDoc.asXML();
}

Value

Script

List strippedlist = new ArrayList();
for (Iterator i = stringlist.iterator(); i.hasNext();) {
	String item = (String) i.next();
	if (strippedlist.contains(item) == false) {
		strippedlist.add(item);
	}
}

Script

String seperatorString = "\n";
if (seperator != void) {
	seperatorString = seperator;
}
StringBuffer sb = new StringBuffer();
for (Iterator i = stringlist.iterator(); i.hasNext();) {
	String item = (String) i.next();
	sb.append(item);
	if (i.hasNext()) {
		sb.append(seperatorString);
	}
}
concatenated = sb.toString();