Sequencing full-length cDNA clones is the most robust way to annotate genes.
Recently, with the advent of new-generation sequencing instruments such
as Illumina Genome Analyzer (Illumina GA) and AppliedBiosystem SOLiD, genome-wide
comprehensive approaches including RNA-sequencing succeeded in capturing
both expressed regions and expression levels of genes without cloning individual
mRNAs. However, RNA expression level varies widely at least four orders
of magnitude, and thus it is impossible to robustly capture lowly expressed
transcripts without cloning or other nomalization techniques. Moreover,
overlapping transcripts (alternative splicing variants, alternative transcription
start/end sites) are difficult to distinguish by RNA sequencing unless
splicing junctions are previously known.
MuSICA 2 aims at shotgun sequencing ~800 full-length cDNA clones using
one flow-cell lane of Illumina GA. First, target full-length cDNA clones
are amplified by PCR, and then they are mixed by equal volumes regardless
of PCR efficiency. The mixture is nebulized to yield a shotgun library,
which is then sequenced by Illumina GA using one flow-cell lane. Finally,
MuSICA 2 assembles the shotgun reads and Sanger reads from both ends of
the cDNA clones (recommended; at least either end is required to associate
the output contigs with individual clones), and reconstructs their cDNA
sequences with the help of the reference genome sequence. The key feature
of MuSICA 2 is that it assembles shotgun reads in a hybrid way of de novo assembly and reference assembly in which shotgun reads are aligned against
the reference genome. The hybrid strategy benefits both from the advantages
of de novo assembly and reference assembly.
Since MuSICA 2 was implemented in Perl, you do not have to compile source
codes. Download the source code from the link below. Please see README.TXT
in the archive for installation and usage. See below for prerequisites and manual.
Velvet and BLAT are prerequisites for MuSICA 2. You have to install and
configure them before you run MuSICA 2. Velvet can be replaced with any
other de novo short-read assembler, but we recommend use of Velvet for the time being
because our benchmark test showed that it was the best assembler that works with MuSICA 2.
Ver 2.1.1, 14 Jul 2010: Small changes conforming with qmake standard use.
Ver 2.1.0, 6 Mar 2009: Significantly reengineered due to a drastic change
in the algorithm.
Ver 1.1.0, 7 Jun 2008: Initial Release
|Download Suppelmentary Data
Raw data used in the above publication can be downloaded from the link
below. The same data in SRF format is also deposited in Short Read Archive.
The raw data below is not required to run MuSICA2 but provided for those
interested (probably tool developers). If you need other data, please contact
to the authors.
- Shotgun reads for Library 1 (seq/prb/qhg/sig2, tar.bz2 format), Intensity file for Library 1 (int/nse, tar.bz2 format).
- Shotgun reads for Library 1+2 (seq/prb/qhg/sig2, tar.bz2 format), Intensity file for Library 1+2 (int/nse, tar.bz2 format).
- Shotgun reads for Library 3 (seq/prb/qhg/sig2, tar.bz2 format), Intensity file for Library 3 (int/nse, tar.bz2 format).
- Reference full-length cDNA sequences (FASTA, tar.bz2 format).
- Sanger reads from both ends of the full-length cDNA clones (FASTA, tar.bz2 format). Note that vector sequences were already trimmed.
- Reginaldo Kuroshu (@cb.k.u-tokyo.ac.jp)
- Masahiro Kasahara (@cb.k.u-tokyo.ac.jp)
- Yutaka Suzuki (@k.u-tokyo.ac.jp)
- Shin-ichi Morishita (@cb.k.u-tokyo.ac.jp)
These releases are provided for archival purposes only. We do not support old releases.