David Jaffe’s team develops genome assembly methods at the Broad Institute. Their goal is to couple optimal laboratory and computational methods to yield the best possible genome assemblies. They developed the widely used algorithms ARACHNE, ALLPATHS-LG, and now DISCOVAR. DISCOVAR starts from a low-cost data type, 250 base reads from a single PCR-free library. From these data the algorithm constructs an assembly graph that has long-range contiguity (100 kb for human genomes) and fully captures polymorphism. Direct visualization of these graphs reveals important features including structural variation, in both normal and cancer samples.