Structural variant calling: the long and the short of it.
Mahmoud M, Gobet N, Cruz-Dávalos DI, Mounier N, Dessimoz C, Sedlazeck FJ. · 2019
Read Structural variant calling: the long and the short of it. by Mahmoud M, Gobet N, Cruz-Dávalos DI, Mounier N, Dessimoz C, Sedlazeck FJ. (2019) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc6868818.txt.
Introduction
Mahmoud and co-authors review methods for finding structural variants, in Genome Biology. Structural variants are large changes in DNA, from chunks deleted or duplicated to whole segments flipped or moved. They shape disease risk, gene control, and the differences between peoples and species. Finding them reliably, though, remains hard.
Different techniques catch different slices of the problem. Short reads spot small events but blur large ones; long reads reach further at higher cost. The review sorts approaches by variant class, size range, and whether one or both gene copies carry the change. It covers combined methods that merge several signal types. The goal is a practical guide for designing a study.
Catalog record
| CREATOR | Mahmoud M, Gobet N, Cruz-Dávalos DI, Mounier N, Dessimoz C, Sedlazeck FJ. |
|---|---|
| YEAR | 2019 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 67 KB |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Structural variant calling: the long and the short of it By Medhat Mahmoud, Nastassia Gobet, Diana Ivette Cruz-Dávalos, Ninon Mounier, Christophe Dessimoz, Fritz J. Sedlazeck Published in: Genome Biology DOI: 10.1186/s13059-019-1828-7 ABSTRACT Recent research into structural variants (SVs) has established their importance to medicine and molecular biology, elucidating their role in various diseases, regulation of gene expression, ethnic diversity, and large-scale chromosome evolution—giving rise to the differences within populations and among species. Nevertheless, characterizing SVs and determining the optimal approach for a given experimental design remains a computational and scientific challenge. Multiple approaches have emerged to target various SV classes, zygosities, and size ranges. Here, we review these approaches with respect to their ability to infer SVs across the full spectrum of large, complex variations and present computational methods for each approach. INTRODUCTION Structural variants (SVs) are large genomic alterations, where large is typically (and somewhat arbitrarily) defined as encompassing at least 50 bp. These genomic variants are typically classified as deletions, duplications, insertions, inversions, and translocations describing different combinations of DNA gains, losses, or rearrangements [1–3]. Copy number variations (CNVs) are a particular subtype of SVs mainly represented by deletions and duplications (reviewed in Carvalho and Lupski [4]). SVs are typically described as single events, although more complex scenarios involving combinations of SV types exist [5, 6]. Chromothripsis, which is a large and complex combination of rearrangements reported in cancer [7], is an example. While the average genomic variation between two humans is 0.1% in terms of single nucleotide variants (SNVs), when taking SVs into account, this increases to 1.5% [8]. In particular, telomeric regions are affected by a higher rate of SVs [9]. SVs can have a pronounced phenotypic impact—disrupting gene function and regulation or modifying gene dosage. Multiple studies have highlighted their role in functional changes across populations [1, 10, 11] and species [12]. Their importance in medicine and molecular biology has been highlighted by multiple recent studies. For instance, in neurological diseases, SVs have been often discussed based on ATTCC repeat extensions in Parkinson [13] or CAG expansions in Huntington disease [14]. Furthermore, a retrotransposon insertion in an intron of the TAF1 gene has been associated with early stages of linked dystonia-parkinsonism disease [15]. In cancer, different types of SVs have been highlighted as causing various types of dysfunction: (i) deletions or rearrangements truncating genes [16]; (ii) amplification of genes leading to overexpression, for example, due to homologous recombination (HR) that leads to an inactivation of BRCA1 and BRCA2 [17, 18]; (iii) gene fusions, such as Her2-positive SKBR3 breast cancer that combines multiple genes across chromosomes [19]; and (iv) alteration of the location of gene regulatory elements, causing changes in the gene expression [4, 20]. In Mendelian studies, multiple diseases have been associated with deletions or duplications of genic regions. For example, three complex SVs affecting ARID1B (Coffin-Siris syndrome), HNRNPU (hypotonia), and CDKL5 (early infantile epileptic encephalopathy is a severe intellectual disability and Rett-like features) have been reported [21]. Another more recent study showed the complexity of these CNVs and an increase in mutation rates for Potocki-Lupski and Smith-Magenis syndrome [22]. SVs are also playing an essential role in plants including having a direct phenotypic impact [23]. For example, SVs play important roles in tolerance for multiple plants: (i) in maize, a tandem triplication over the AMTE1 genes is reported to be associated with aluminum resistance [24]; (ii) an amplification of Bot1 plays an important role in boron toxicity in barley [25]; and (iii) for weeds, a tolerance against the herbicide glyphosate based on amplification of EPSPS has been reported in response to extensive use of glyphosate [26]. Other SVs have a positive impact on fruit yield and quality. For example, a transposon insertion near Ruby, a MYB transcriptional activator, leads to the increase of anthocyanin concentration in blood orange compared to pumelo and mandarin [27]. In tomatoes, a transposon insertion in JOINTLESS2 (J2) results in undesirable branching of flower-bearing shoots (inflorescences) in genetic backgrounds that also carry a cryptic variant for the close homolog enhancer of J2. This combination results in excessive flower production. However, an additional tandem duplication in fresh-market breeding lines across this region leads to a threshold of correctly spliced product and thus to a healthy phenotype with higher fruit yield [28]. Despite all these evidences of the importance of SVs, they have been largely understudied, compared to SNVs, because they are much more difficult to identify. In principle, taken individually, each type of SV induces a distinctive pattern in mapping reads that can be used to infer the underlying mutation. For example, a deletion forms a lack of a sequence and thus a gap in the alignment of the sample relative to a reference (Fig. 1). However, in practice, it is much more complicated. First, sequencing and mapping errors blur the patterns. Indeed, in contrast to SNVs and smaller insertions and deletions, SVs can cover a large portion of a read or even be larger than the read length—which complicates mapping [5]. Second, the patterns induced by the different SV types can be very similar. For example, it is often hard to distinguish tandem duplications from novel insertions for genomic alignments (Fig. 1). Finally, multiple SVs can overlap or be nested, giving rise to much more complex mapping patterns than when considered individually [5, 20]. Such complex patterns may preclude mapping altogether, forcing researchers to assemble each genomic sample de novo—a difficult and more costly task with conventional sequencing. Fig. 1Comparison between de novo assembly, short-read and long-read mapping approaches to identify structural variants. For de novo assembly approaches, the relative positions of the segments in the dot plot indicate the type and size of the SV. For short-read-based mapping approaches, paired-end (red) and split reads (purple) are typically used to decipher the type size and location. In addition, the coverage can be used to improve the detection of deletions and duplications. Long-read-based mapping approaches typically leverage the alignment patterns of long reads (green) to detect the different types of SVs However, great strides have recently been made, thanks to technological and methodological developments. The advent of long-read sequencing technology, in particular, Pacific Biosciences (PacBio) and Oxford Nanopore technologies (ONT), makes it possible to produce reads of several thousand base pairs, even reaching up to 2 Mbp for Oxford Nanopore [29]. Furthermore, as we shall review in more detail below, technologies such as linked reads (e.g., 10x Genomics), optical mapping, and Strand-Seq have also been developed to improve the quality of assemblies and/or SV calling. Long reads help to increase the detection of SVs as they considerably ease de novo genome assembly and mapping. Nevertheless, the increased length and the higher error rate of emerging long-read technologies can pose new methodological challenges. Complementary to long reads, another noteworthy development has been the repurposing of transcriptomics (RNA-Seq) to detect SVs—in particular, rearrangements. Indeed, by identifying apparent RNA fusions, which are thus inherently transcribed, it is possible to focus on SVs with potential functional implications. Lastly, recent progress in benchmarking is greatly improving our understanding of the strengths and weaknesses of each approach. Current efforts such as Genome in the Bottle [30] and the FDA-led initiative SEQC2 (https://www.accessdata.fda.gov/scripts/fdatrack/view/track_project.cfm?program=nctr&id=NCTR-DBB-PM-SEQC2-Phase-II) aim at better characterizing false