The following java code reads the flat file nodes.bmp and print the common ancestor between two taxin-ids.

import java.io.*;
import java.util.*;
import java.util.regex.*;
public class Biostar10350
    {
    private Map<Integer,Integer> id2parent=new HashMap<Integer,Integer>(791000);
    /** get all ancestor of a given taxon id*/
    private List<Integer> lineage(int id)
        {
        if(!id2parent.containsKey(id)) throw new IllegalArgumentException();
        LinkedList<Integer> L=new LinkedList<Integer>();
        for(;;)
            {
            L.addFirst(id);
            Integer parent=id2parent.get(id);
            if(parent==null || parent.equals(id)) break;
            id=parent;
            }
        return L;
        }

    private int run(String filename,int id1,int id2) throws Exception
        {
        Pattern pipe=Pattern.compile("[\\|]");
        BufferedReader in=new BufferedReader(new FileReader(filename));
        String line;
        while((line=in.readLine())!=null)
            {
            String tokens[]=pipe.split(line,3);
            Integer tax_id=Integer.parseInt(tokens[0].trim());
            Integer parent_id=Integer.parseInt(tokens[1].trim());
            id2parent.put(tax_id,parent_id);
            }
        in.close();
        List<Integer> L1= lineage(id1);
        List<Integer> L2= lineage(id2);
        int index=-1;
        while(index+1 < L1.size()  &&
            index+1 < L2.size() &&
            L1.get(index+1).equals(L2.get(index+1)))
            {
            index++;
            }
        return L1.get(index);
        }

    public static void main(String args[]) throws Exception
        {
        System.out.println("Common ancestor is taxon-id:"+
            new Biostar10350().run(
                args[0],
                Integer.parseInt(args[1]),
                Integer.parseInt(args[2])) );
        }
    }

Usage:

$ javac Biostar10350.java 
$ java Biostar10350 nodes.dmp  9606 10090
Common ancestor is taxon-id:314146

Hi,

I just published a blog post about how you can achieve that (also extended to a set of nodes with an arbitrary length) with Bio4j

http://blog.bio4j.com/2012/02/finding-the-lowest-common-ancestor-of-a-set-of-ncbi-taxonomy-nodes-with-bio4j/

Hope it's useful ;)

Pablo