For the detection of large genetic variations (structural variants), short-read sequencing technology proves to be insufficient. To maintain the high resolution of short reads, expensive long-read sequencing is often applied. A less costly option is optical mapping, which uses base motif labeling with fluorescence microscopy to obtain genome data. The produced molecules are three to six times longer than long-reads on average but do not contain nucleotide information; the data consists of label positions of the base motif instead. This combined with a high error rate proves the alignment and assembly of molecules necessary for variant detection challenging. As part of a project collaborating with the university clinic, we analyze samples of leukemia patients from pediatric oncology and develop new algorithms to detect structural variants in optical mapping data and investigate their relation to the onset and progression of cancer.