HiSeq Analysis Software FAQs

Collapse All

|

Expand All

  • General


  • HiSeq Analysis Software (HAS) provides rapid, easy alignment and variant calling for whole human genomes or libraries prepared with the Nextera Rapid Capture Exome Enrichment kit

    • For human whole-genome sequencing (WGS), HAS features the Isaac analysis workflow. Isaac provides a 4–6x speed increase over existing methods.
    • For Nextera Rapid Capture analysis, the BWA alignment and GATK variant calling methods are used.

    The software is command-line, or you can use a graphical user interface package called Analysis Visual Controller Software (AVC).

    No, the HiSeq Analysis Software provides analysis of libraries prepared with the Nextera Rapid Capture exome enrichment kit and human Whole Genome Sequencing using only the hg19 genome as a reference. CASAVA provides analysis of additional applications such as RNA sequencing, Exome sequencing, targeted resequencing, and Whole Genome Sequencing using a more extended set of reference genomes available on the Illumina iGenomes page

    Isaac has slightly lower sensitivity and specificity, but is much faster than BWA/GATK.

     

    Conflicts

    Conflict Rate

    Sensitivity

    Isaac

    6318

    0.139%

    94.5%

    BWA+GATK

    5315

    0.126%

    95.8%

    The open source version of Isaac includes the component algorithms for the Isaac aligner and variant caller. It is intended for developers, and is not commercially supported. Instead, it is provided as is under Illumina Open Source Software License.

    For most Illumina customers, we recommend using Isaac as part of the HiSeq Analysis Software (HAS) package. HAS provides rapid and easy alignment and variant calling for Whole Human Genomes (using the Isaac component algorithms) or Nextera Rapid Capture Exome Enrichment libraries (using the BWA/GATK component algorithms). The software is available through a command line interface or a graphical user interface package called Analysis Visual Controller Software (AVC).

    HAS is freely available, easy to install (rpm), and commercially supported. For more information, HiSeq Analysis Software.

    The Isaac variant caller is based on a Bayesian model and assigns probabilities to different possible variant calls, so no specific minimum exists.

    Isaac has reliably detected 10 bp continuous indels over the length of the read.

    The Isaac aligner aligns reads by first identifying a small but complete set of relevant candidate mapping positions. The Isaac aligner begins with a seed-based search, using 32-mers as seeds. The initial single-seed search is followed by a multi-seed only for the reads that couldn't be placed unambiguously with a single seed. Speed up is achieved by sorting the reference index by the 32-mers. Improvement to accuracy is achieved by flagging of all the ambiguous reference positions in the index.

    Following the seed-based search, selection of the best mapping among all the candidates is performed. For paired-end data sets, all mappings where only one end is aligned (orphan mappings) trigger a local search to find additional mapping candidates (shadow mappings) in the neighborhood defined by the expected minimum and maximum insert size. After optional trimming of low quality 3' ends and adaptor sequences, the possible mapping positions of each fragments are compared, taking into account pair-end information when available, possible gaps (using a banded Smith-Waterman gap aligner) and possible shadows. The selection is based on the Smith-Waterman score (using BWA, ELAND or user-defined scores) and on the log-probability of each mapping. The main speed-up comes from a parallel implementation of the gap aligner (using the SSE2 instruction set) and a shadow aligner optimized for short inserts. Further improvements could be achieved with AVX. The gapped alignment could be delegated to a coprocessor (e.g. Xeon Phi or GPU), however it is unclear if the benefit of large-scale parallelization would outweigh the cost of transferring the data between host and coprocessor.

    Following alignment the fragments are sorted. Major speed-up in the sorting speed-up comes from efficient binning of the selected mappings, which greatly simplifies the sorting. Further analysis is performed to identify duplicates and optionally to re-align indels.

    Source code for the Isaac algorithms is availble here: github.com/sequencing.

    The software is released is under Illumina Open Source Software License available on github.com/sequencing/licenses/. The license terms are intended to make the software accessible to a variety of bioinformatics developers, computational biologists, and other users in the research community, and to encourage community development of the software. For questions about the license, contact Isaac-admin@illumina.com.

    A peer-reviewed application note is available on the Bioinformatics website:

    Raczy C, Petrovski R, Saunders CT, Chorny I, Kruglyak S, et al. (2013) Isaac: Ultra-fast whole genome secondary analysis on Illumina sequencing platforms. Bioinformatics 10.1093/bioinformatics/btt314