Simulate differential ChIP-seq regions

DCSsim simulates differential ChIP-seq data for two samples (e.g. treatment and control). The results are sequence reads in fasta format for the two samples with n (user defined) replicates plus a report as pdf. DCSsim supports multiple threads and is able to work in batches to account for limited memory. We strongly encourage new users to start with the two parameter simulation scripts test-peak-shape.py and test-fragment-count-distribution.py. These can be applied to test the parameters for sampling without creating reads but output diverse histograms and plots.

DCSsim

DCSsim.py <FASTA> <BED> <INT> [options]

Based on the provided <FASTA> sequence differential peaks (DPs) are simulated, restricted to if --is_white_list or constrained by the specified <BED>-file. <INT> defines the number of simulated replicates of the two samples.

Options:

-c, --chrom Chromosome used for simulation, default='chr1',

-d, --domain-counts Number of domains/clusters with DPs, default=1000

-l, --length Read length, default=50

-p, --prefix Prefix for output files, default='sim'

--prot-size Protein size (this parameter limits the fragment shifts), default=150

--prot-count-n n of negative binomial distribution for protein counts, default=1

--prot-count-p p of negative binomial distribution for protein counts, default=0.9

--prot-dist-muno-mean Means of multivariate normal distribution for distances between proteins, separator: ',' eg. "300, 1800", default=300, 1800, 900

--prot-dist-muno-cov Covariance of multivariate normal distribution for distances between proteins, separator: ',' and ';' eg. "1000,0;0,5000", default=[[1000,0,0],[0,5000,0],[0,0,5000]]

--frag-len-mean Set mean of fragments' length, default=200

--frag-len-dev Set deviation of fragments' length, default=20

--frag-len-max Set maximum of fragments' length, default=1000

--frag-count-sh Shape of gamma distribution for fragment counts, default=2.2

--frag-count-sc Scale of gamma distribution for fragment counts, default=20.1

--frag-count-op Probability for fragment counts being outliers, default=0.01

--frag-count-om Mean of lognormal distribution for fragment counts of outliers, default=6.0

--frag-count-os Sigma of lognormal distribution for fragment counts of outliers, default=0.5

--frag-count-scaling Scaling of fragment distribution, no scaling, scaling of beta result based on fragment counts (with exponential distribution) or scaling of fragment counts based on beta result (with Laplace distribution): none, frag, beta, default="none"

--frag-count-lp-scale Scale for Laplace distribution if frag-count-scaling is frag, default=0.1

--frag-count-ex-loc Loc for exponential distribution if frag-count-scaling is beta, default=10

--frag-count-ex-scale Scale for exponential distribution if frag-count-scaling is beta, default=100

--frag-dist-on Use multivariate normal distribution for fragment shifts to create peak shapes, shifts are limited by prot-size. The final shift is: postion of peak - prot_size + sampling from distribution, default=False

--frag-dist-prob, dest Probability for each of the multivariate normal distributions to be chosen, default=[0.5, 0.5]

--frag-dist-muno-mean Means of multivariate normal distribution for the shifts of fragments, separator: ',' eg. "300, 1800", default=[20, 100]

--frag-dist-muno-cov Covariance of multivariate normal distribution for the shifts of fragments, separator: ',' and ';' eg. "1000,0;0,5000", default=[[100,0],[0,500]]

--beta Alpha and Beta of Beta-distribution, default=[0.5, 0.5]

--dp-thres Threshold of reads/fragments to define a DB peak, default=0.6

-m, --min-counts Minimum number of reads/fragments for a DB peak, default=25

-s, --skewness Variance between replicates (the higher, the less variance), default=10

--back-avg Average background coverage for noise, default=0.25

--back-res Resolution for ChIP-seq noise estimates (also used for spike-in and input), default=1000

--back-c Gamma distribution scale (theta) for noise model (also used for spike-in and input), default=20

--back-s Gamma distribution shape (k) for noise model (also used for spike-in and input), default=1

--is-white-list Set provided bed-file to white-list and alow only DPs in these regions ", default=False

--no-input Create no input/control-fasta per sample, default=False

--no-fasta Do not create fasta files, only the report will be created, default=False

--no-report Do not create pdf report, default=False

--no-noise Do not add noise to the simulated reads/peaks, default=False

--spike-in Add spike-in reads from a defined reference fasta, default=False

--si-fasta Path to reference fasta file for spike in, default='spike_in.fasta'

--si-chrom Chromosome of spike-in reference fasta file to simulate, default='chr1'

--si-cov Background coverage for spike-in, default=0.25

-t, --threads Number of threads to use, default=1

--batch-size Number of domains/clusters calculated in one batch to limit memory usage, default=10000

test-fragement-count-distribution

test-fragement-count-distribution.py <INT> <INT> [options]

Creates a histogram of <INT> protein interactions sites of <INT> replicates for two samples, for all samples, for sample1 and sample2 and creates an MA-plot based on the supplied beta-values