FAQ

Frequently asked questions regarding to installation and usage of FuSViz.

1. How to prepare input format for FuSViz?

  • FuSViz accepts three types of data as input: SV calls from RNA-seq/DNA-seq data and mutation profile (i.e. small variants). Here, an in-house script SV_standard.pl (see its usage in Installation section) is provided to convert raw SVs from various callers to a uniform format, then aggregate calls of multiple samples together. In terms of SVs from RNA-seq data, many callers have been developed and there lacks of a standardization format for the output. In the processing of SV_standard.pl, a few common features (e.g. gene name, breakpoint position, sample id, read support for SV event) shared in the output of most callers are extracted and standardized as an input format of FuSViz. In contrast, most SV callers uing DNA-seq data output a standard VCF format, which can be converted to bedpe format with some key fields kept for FuSViz analysis. Now, output format of 10 SV callers (STAR-fusion, Arriba, deFuse, FusionCatcher and Dragen for RNA-seq; Menta, Lumpy, Delly, SvABA and Gridss for DNA-seq) are well supported by SV_standard.pl, and a support for more callers will be extended on basis of user’s requirement and feedback in the further.

2. What are differences between SV calls from DNA-seq and RNA-seq, and their respective characteristics? Why to combine them?

  • Genomic SV at DNA level represents various types of large variations (e.g. deletions, insertions, duplications, inversions and translocations) that are related to the break, link, recombination and amplification of DNA strand. When it occurs in genic regions, a hybrid gene is formed (so-called “fusion gene”). In case of breakpoint in non-coding regions, duplications and deletions can lead to an alteration of genomic regulatory elements (e.g. enhancers or TADs). One common feature of SV rearrangement is to break up DNA sequence, which disrupts a proper transcription or translation of tumor suppressor gene, finally giving rise to a reduced activation or function loss of the gene.

  • In comparison, RNA-seq data is used to address whether a fusion gene can transcribe to a fusion transcript and produce a putative chimeric protein with an in-frame outcome. Although breakpoints of fusion genes usually occur in intronic regions, the observed breakpoints of most fusion transcripts coincide with exon boundary due to RNA splicing mechanism, from which some breakpoint pairs show a higher prevalence than others. One important biological consequence of fusion transcript is an activation of oncogene, in particular for a oncogenic downstream partner. In addition, a few SV types, e.g. read-through and trans-splicing, are exclusively detected at RNA level which represents a event of RNA-mediated regulatory mechanism.

  • In short, it is necessary to combine features of SV calls from DNA-seq and RNA-seq data together to interpret and visualize SVs associated with oncological-relevant genes in context of cancer genomic and transcriptomic annotations.

3. Are SV calls from the third-generation sequencing technology (e.g. PacBio or Nanopore platform) supported by FuSViz.

  • Although FuSViz is designed to visualize and interpret aggregated SVs called from short read sequencing data (main for the Illumina platform), it provides a support for SV calls from long read sequencing data (e.g., Nanopore or PacBio technology) as well. To our knowledge, most SV tools using long read sequencing data output raw calls into a standard VCF file with a similar format as those from short read data. Users just need to convert SV calls to a format compatible with FuSViz requirement (i.e., bedpe or bedpe-like format).

4. If I am only interested in copy number aberrations (CNAs), how could I make a filtration on SVs.

  • In general, CNAs represent abnormal amplifications and deletions of intra-chromosomal genomic regions. They are a sub-group of total SVs, which are tagged as DUP and DEL in the input format of FuSViz. If users only focus on an analysis of CNAs, please click on the button Load and refresh DNA SV Track in seg in Linear plot module. In addition, users can use SV_type Select box to filter out other types of SVs for a customized analysis.

5. Can it be possible to utilize FuSViz for analysis of a single sample and ‘tumor-normal’ pairwise samples?

  • Users can utilize name or Sample in Select box to focus on their interested samples. Moreover, a few functionalities are provided to enhance the single sample analysis by allowing import of a read alignment file in linear module (e.g. SV quality control in Appendix section) and make a fusion plot including read converage of partner genes in two-way module under a CLI mode. However, FuSViz is not designed for a comparison analysis between “tumor” and “normal” pairs except for exploring genetic differences in read alignment.

6. Is it possible for FuSViz to analyze SV calls from Non-human organisms?

  • FuSViz is a dedicated tool intended for interpretation of genomic/transcriptomic structural variations (SVs) in the context of human cancer, and hence has limited focus on serving datasets from other organisms. But, we have extended its functionality to the model organism mouse, given that mouse is frequently used to establish patient-derived xenograft (PDX) models for experimental investigations of molecularly targeted cancer therapies.

7. Do SV calls in FuSViz input file include genotype information?

  • No, there are several reasons for it:

    • Firstly, genotype information of SVs is not a common feature presented in raw output of most callers.

    • Secondly, the precision of genotype information highly depends on the type of SVs. In general, duplication and deletion are genotyped relatively accurate. Genotype of insertion and inversion can be inferred but with less precision. Most tools have a limited capacity to genotype translocation.

    • Thirdly, it is challenging to make a consensus on SV genotypes in a merge process if discordance from different callers is present.

8. Does FuSViz have a functionality to allow alignment file (BAM) as input?

  • Although the main purpose of FuSViz is used to visualize and interpret SVs of multiple samples, users can load read alignments in the file upload box of Linear module or using Two-way module for a single sample analysis (see Appendix - how to plot read coverage for a customized analysis of one specific sample using alignment file, either from RNA-seq/DNA-seq or both).