Table 1 Structural variant (SV) callers in use at clinical sites.

From: Commonalities across computational workflows for uncovering explanatory variants in undiagnosed cases

 

BaylorSeq

BCM

Duke/Columbia

Harvard

Miami

NIH

PacificNW

Stanford

UCLA

Utah

Vanderbilt

WUSTL

Find SVs from sequencing reads

  Mantaa

  ExpansionHunter

 

  

   

 

  GATKb

 

 

 

  

   

  LUMPY

    

   

 

  CNVnator

    

   

  

  RUFUS

 

       

  

  CNVkit

       

   

  BreakDancer

    

      

  Illumina DRAGEN depth-based CNV caller

           

  SvABA: SV/indel Analysis by Assembly

   

        

  CoNIFERc

      

     

 ERDS: estimation by reads depth w/ SNVs

        

   

  BreakSeq2

    

       

  DELLY2

    

       

Jointly call and/or genotype SVs

  smoove

      

  

 

  SVTyper

    

    

 

Annotate SVs

  AnnotSV

 

    

 

  gnomAD-SV

    

      

  duphold

         

 

Run or combine output from other tools

  XHMM

 

 

 

      

  SURVIVOR

    

      

  Parliament2

    

       
  1. ■ Tool called directly. □ Tool called indirectly (e.g., by a wrapper).
  2. Each SV calling tool identifies subsets of SVs by type or other factors, and so in practice, the output of multiple methods must typically be combined and considered together. Wrapper tools that automatically call and combine results from multiple other SV detection methods improve the efficiency of this process. Duke/Columbia, NIH, Stanford, and Vanderbilt only use SV calling tools in specific cases or contexts rather than as part of their regular pipelines. Tool citations are listed in Extended Data Table 1.
  3. CNV copy-number variant, SNV single-nucleotide variant.
  4. aManta is used by BaylorSeq to generate putative SV calls, which are then shared with the clinical sites.
  5. bThe two functions from GATK used are GermlineCNVCaller and DepthOfCoverage (DoC); the latter is used to detect exonic deletions or duplications.
  6. cIn contrast to other tools, CoNIFER runs on exome sequencing (ES) data rather than genome sequencing (GS) data.