A pair of researchers from the University of Luxembourg introduces an in silico reference model-based method for finding large-scale structural variants in cancer genomes. The approach uses breakpoint patterns to appropriately place stray or mismatched reads in the context of many small reference sequences generated in silico from patient samples and/or cell lines, offering a look at structural variants in the process, authors of the study note. "By using these models to align unmapped reads," they write, "we show that our method can help to identify large-scale variations that have been previously missed."
British Columbia Cancer Agency and University of British Columbia researcher Marco Marra and colleagues provide a peek at the approach they used to profile microRNAs in more than 11,000 tumor samples from almost three dozen cancer types tested for the Cancer Genome Atlas project. The team detailed steps taken along the path from library production to data submission, providing information for those interested in performing similar miRNA assessments or comparing their data to the miRNA data produced for TCGA. The study's authors note that "[w]hile we focus here on applying the pipeline to large-scale cancer datasets in the context of TCGA, the modular pipeline can run on any SAM-format read alignment file generated for any species that has a reference genome and annotated miRNAs, accessing annotations from a database or from user-created text files."
A University of Massachusetts team shares findings from a study aimed at uncovering structural variants in the genomes of almost 100 cancer samples. Using a local assembly-based algorithm they developed called laSV, the researchers scrutinized matched tumor and normal genome sequences generated by TCGA for 97 cases representing half-a-dozen cancer types. In tumor samples from individuals with acute myeloid leukemia, for example, they found that somatic structural variants often seemed to arise as a consequence of non-allelic homologous recombination. On the other hand, the structural alterations appearing in solid tumor types appeared more apt to stem from mechanisms such as non-homologous end joining or microhomology end joining.