Bioinformatics / detection of periodic patterns in gene expression data
C R HTML PostScript TeX CSS Other
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
code_C
code_R
cosopt.wiki
.RData
.Rhistory
README.md
README.md~

README.md

cosopt

Bioinformatics / detection of periodic patterns in gene expression data

Release Info

Realsed in April 6, 2015. Updated in January 22, 2016. Validated on Mac OS 10.8.5 and R 3.2.0.

##Introduction COSOPT is a statistical method to distinguish a period and phase for a transcript in time-series gene expression analysis. The source code is provided by C language and R.

IMPORTANT: This COSOPT has a small difference from the original COSOPT described in Straume et al. 2004. In the original paper, SD within replicates of a probe was used to generate surrogates with white noise, however we fixed SD to 0.1 in our analysis. If you want to change SD values yourself in the R package, please modify variable as follows.

R version

Installation

You can find a compressed package such as "cosopt_0.3.tar.gz" in code_R/ directory. Just use install.packages command as below:

> install.packages("cosopt_0.3.1.tar.gz",repos = NULL, type = "source")

Running

We recommend to run COSOPT against a small number of samples on R.

> timepoints <- c(0,8,16,24,32,40,48)     # time points of gene expression data
> data <- c(-1,0,1,0,-1,0,1)  # signal intensity of each probe in gene expression data
> sigma <- c(0.1,0.1,0.1,0.1,0.1,0.1,0.1)  # sigma (standard deviation of each probe in the original source code) 
> cosopt(data,sigma,timepoints,plotting=TRUE) # run COSOPT with plotting image
gene=1 timecourse=7
plotting=1
minimum_freq=0.031250

A period of this gene is T= 1/f = 1/0.031250 = 32(h).

C version

The original source code is written in C. You can find it from code_C/.

Running COSOPT by C with MPI environment

Here is MPI_COSOPT which is optimized for parallel computation, MPI. Running MPI_COSOPT requires knowledge of parallel computing. You can also run MPI_COSOPT on SunGridEngine(SGE/UGE).

Supplementary data

All supplementary data are available from the github wiki page or our paper.

Citation

Please cite the following paper if you used this program.

  • Hiromi Matsumae, Ryosuke Ishiwata, Toshifumi Minamoto, Norio Ishida, Soichi Ogishima, and Hiroshi Tanaka. "Detection of periodic patterns in microarray data reveals novel oscillating transcripts of biological rhythms in Ciona intestinalis" 2015. Artificial Life and Robotics, Springer. doi:10.​1007/​s10015-015-0237-6 Link to the paper: http://link.springer.com/article/10.1007/s10015-015-0237-6

Contact

Please report a bug and ask a question to the following contacts.

  • Hiromi Matsumae: hiromi [dot] matsumae [atmark] gmail [dot] com
  • Ryosuke Ishiwata: ishiwata [atmark] phys [dot] cs [dot] is [dot] nagoya-u [dot] ac [dot] jp