DICEseq (Dynamic Isoform spliCing Estimator via sequencing data) estimates the dynamics of isoform proportions jointly from time series RNA-seq experiments. DICEseq is a Bayesian method based on a mixture model whose mixing proportions represent isoform ratios. It incorporates the correlations from the temporal structure, by coupling the isoform proportions at different times through a latent Gaussian process (GP).
DICEseq provides following functions through command line:
diceseq: estimate the isoform proportions and FPKM for time series data jointly, or for a single time point.dice-count: fetch reads counts for entile gene, or specific reads counts, e.g. junction reads, for genes with exact one intron. This is special design mainly for yeast.dice-bias: estimate parameters for sequencing bias, including fragment length distribution, reads sequence and position bias parameter. The output file can be directly used for bias correction indiceseq.
In addition, DICEseq package also provides interface of a set of functions and
attributes as an object-oriented python module. Therefore, you could use some
of the module e.g., SampleFile to visualize the samples in gzip file in a
Gaussian process way, or BiasFile to visualize the bias parameters. Also,
the gtf_utils provides a set of ways to load gtf file, choose the genes, or
customize the coordinates of exons and introns, add and remove of specific
transcripts.
Environment and installation:
DICEseq was initially developed in Python 2 environment, hence best to be used in Py2 environment. By using Anaconda platform, no matter Py2 or Py 3, it is easy to set up a conda environment with Py2, for example by following commond:
conda create -n dicePy2 python=2.7 numpy==1.15.4 scipy==1.1.0 matplotlib==2.2.3 pysam==0.15.2
source activate dicePy2Once you are in a Python 2 environment, there are usually two ways to isntall a package:
pip install diceseq- or download this repository, and type
python setup.py install. - You may need to add
--userif you don't have the root permission for that environment.
Arguments
- Type command line
diceseq -h
See the documentation on how to install, to use, to find the annotation data etc.
Yuanhua Huang and Guido Sanguinetti. Statistical modeling of isoform splicing dynamics from RNA-seq time series data. Bioinformatics, 2016, 32(19): 2965-2972.