I've done some clustering based on the DCJ (Double-Cut-or-Join) distances reported by the Mauve alignment program for 10 assemblies vs. a common reference. I used Mauve to re-order each of my assemblies to that reference before running the multiple alignment, and that I took the pairwise DCJ values that Mauve produced from that alignment for clustering.
But my assemblies are in from 30-100+ contigs each. With my data it doesn't appear that the contig number correlates directly with the DCJ distances, but I was wondering if anyone could confirm that. A colleague tracked down a very brief summary in an NCBI paper:
"The double cut and join (DCJ) distance (Bergeron et al., 2006) between the assembly and reference is calculated to estimate the effect of misassembly and lack-of-assembly errors (excess contig breaks) on rearrangement distance"
But that is a bit unclear to me. Looking at the referenced article, the crux of DCJ seems like its looking at oriented genes sitting on chromosomes. So if you know the orthologous genes and their orientations/positions on the reference you could calculate the # operations to bring the query assembly geneset in line with the reference geneset.
But that doesn't seem possible for Mauve, which only has my assemblies (no genes). Could someone more familiar with Mauve explain how it calculates DCJ?
Thanks,
John Martin
But my assemblies are in from 30-100+ contigs each. With my data it doesn't appear that the contig number correlates directly with the DCJ distances, but I was wondering if anyone could confirm that. A colleague tracked down a very brief summary in an NCBI paper:
"The double cut and join (DCJ) distance (Bergeron et al., 2006) between the assembly and reference is calculated to estimate the effect of misassembly and lack-of-assembly errors (excess contig breaks) on rearrangement distance"
But that is a bit unclear to me. Looking at the referenced article, the crux of DCJ seems like its looking at oriented genes sitting on chromosomes. So if you know the orthologous genes and their orientations/positions on the reference you could calculate the # operations to bring the query assembly geneset in line with the reference geneset.
But that doesn't seem possible for Mauve, which only has my assemblies (no genes). Could someone more familiar with Mauve explain how it calculates DCJ?
Thanks,
John Martin