PART and PART-up

 

Requirements:


perl, JAVA, R and the R poibin package

(http://cran.r-project.org/web/packages/poibin/index.html)


This program has been tested only on Unix and Mac.



How to install:


  1. 1)Download a file

wget  http://www.hgc.jp/~niiyan/PART/part.tar.gz


2) Uncompress the file

tar xvzf part.tar.gz


the resulting  directory contains a README file, the "java" and "data" directories, and the  "part", and "part_up" perl scripts.



How to prepare input data:


PART and PART-up need segmented copy number or LOH data and SNP or CGH array probe information.

As examples, TCGA colorectal cancer LOH data are provided in the data directory.


1) A segment  file

This is a seg format file from the Circular Binary Segmentation algorithm.

(http://www.bioconductor.org/packages/2.3/bioc/html/DNAcopy.html)


2) A  probe information file

This is a tsv file containing probe ids, chromosome ids (1,2..24), coordinates on each chromosome, in the 1st, 2nd and 3rd columns respectively.



How to run:


1) A basic usage is:

perl part.pl  paired.seg probe.tsv

perl part_up.pl  tumor.seg  normal.seg probe.tsv


2) If you change cutoff for aberrations (default: 0.5),

perl part.pl -c 0.6  paired.seg probe.tsv

perl part_up.pl  -c 0.6 tumor.seg  normal.seg probe.tsv


3) If you assumes that aberrations is less than the cutoff,

perl part.pl -l  paired.seg probe.tsv

perl part_up.pl  -l tumor.seg  normal.seg probe.tsv


4) If you want to decrease the number of the counted probes,

perl part.pl -t 1000   paired.seg probe.tsv

perl part_up.pl  -t 1000 tumor.seg  normal.seg probe.tsv

These count one per every 1000 probes.


The result is printed out to the standard output.

Each line contains the probe id, chromosome, coordinate, p-value and q-value.



Reference:


Niida, A and Imoto, S and Shimamura, T and Miyano, S

Statistical model-based testing to evaluate the recurrence of genomic aberrations

Bioinformatics, 2012, 28:i115-i120



Contact:


Atsushi Niida  aniida@ims.u-tokyo.ac.jp