Home

# Extract data quickly from Hi-C files via straw Straw is a fast implementation of reading/dump for .hic files. It is currently primarily written in C++, with support for R and Python. We are working on an all-Python API.

Straw works by reading the .hic file, finding the appropriate matrix and slice of data, and printing the text in sparse upper triangular format.

It is not as fully featured as the Juicebox command line tools Java version; in particular, it doesn't store the pointer data for all the matrices, only the one queried. It also doesn't currently support reading from a URL rather than a local file. Future versions may look more like the Java implementation. Currently only supporting matrices (not vectors).

Please see the below sections on compiling for C++, Python, and R.

For questions, please use the Google Group.

If you use this tool in your work, please cite

Neva C. Durand, James T. Robinson, Muhammad S. Shamim, Ido Machol, Jill P. Mesirov, Eric S. Lander, and Erez Lieberman Aiden. "Juicebox provides a visualization system for Hi-C contact maps with unlimited zoom." Cell Systems 3(1), 2016.

C++

Straw must be compiled with the -lz flag to include the zlib.h library: g++ -lz -o straw main.cpp straw.cpp

Usage: straw <NONE/VC/VC_SQRT/KR> <hicFile> <chr1>[:x1:x2] <chr2>[:y1:y2] <BP/FRAG> <binsize>

Users have also reported needing to use this flag to compile: -std=c++11

R

Straw is compatible with R via the Rcpp library. Usage is very similar to the above. Future versions might use a different signature to call the function.

library(Rcpp)
sourceCpp("straw-R.cpp")
A<-straw_R("NONE /Users/nchernia/Downloads/drosophila.hic arm_2L arm_2L BP 100000")

In the above example, A is a data frame containing the counts information:

> head(A)
        x        y counts
1  800000  4700000     77
2       0        0  41105
3 5500000  8100000     82
4 2700000  8300000     85
5 5700000 17700000     11
6 3000000 16500000     27

Python

Straw can be called from Python by using Boost http://www.boost.org/

The files Jamroot, straw.py, and main-python.cpp are used to hook the C++ to Python. main-python.cpp is essentially the same as main.cpp

1. Get Boost:
   http://www.boost.org/doc/libs/1_61_0/more/getting_started/unix-variants.html#get-boost
2. Compile bjam:
   http://www.boost.org/doc/libs/1_49_0/more/getting_started/unix-variants.html#prepare-to-use-a-boost-library-binary (just do --with-libraries python)
3. Type "bjam" in the directory with the files (assuming you've put boost and bjam in your path).