accurate and fast detection of complex and nested structural variations using long read technologies

Fritz Sedlazeck, Johns Hopkins University, Department of computer science, Baltimore, USA

The impact of structural variations (SVs) is becoming more prominent within a variety of organisms and diseases, especially human cancers. Short-read sequencing has proved invaluable for recognizing copy number variations and other simple SVs, although has been highly limited for detecting most other SVs because of repetitive elements and other limitations of short reads. The advent of long-read technologies, such as PacBio or Nanopore sequencing that now routine produce reads over 10,000bp, offer a more powerful way to detect SVs. However, currently available methods often lack precision and sensitivity when working with highly erroneous reads, especially for more complex or nested SVs.

I will introduce Sniffles and NGM-LR, new fast and accurate methods to map long reads from both PacBio and Oxford Nanopore, and to infer SVs of various sizes and complexity. Working with genuine PacBio and Oxford Nanopore reads with human cancer samples (SKBR3), healthy human samples (Genome in a Bottle), and other species (rice, yeast, E. coli), we show how Sniffles combined with NGM-LR produces the highest quality SV calls available. We further show how these technologies can also reduce the coverage, and therefore cost, required per sample for highly sensitive and specific SV detection. Sniffles and NGM-LR are available open-source at Github, and are already being used by multiple institutes around the world.


Short CV

Fritz Sedlazeck completed his DI in 2008 at the University of Applied Science in Hagenberg. In 2008 he worked at the European Molecular Biology Laboratory (EMBL) in Heidelberg. During his PhD (2008-2012), Fritz focused on the development of novel methods to analyze high throughput sequencing data at the Max F. Perutz Laboratories in Vienna under the supervision of Arndt von Haeseler. After a 2 year PostDoc, he joined the group of Michael C. Schatz at Cold Spring Harbor in New York. Currently, Fritz is a PostDoc at the Johns Hopkins University in Baltimore where he is working on the detection of structural variations, mapping and assembly problems using high-throughput sequencing technologies.