For additional information, including a tutorial and sample data for the epigenotyping procedure, check out the updated documentation.
Stable inheritance of DNA methylation allows creation of epigenotype maps and the study of epiallele inheritance patterns in the absence of genetic variation
-
All python scripts are meant for Python 3.4+
-
All scripts import sys, math, glob, multiprocessing, subprocess, os, bisect, and random (all/most come up python)
-
Many scripts import pandas, numpy, scipy, and/or sklearn
-
To avoid package dependencies, use install anaconda
-
Most R packages can be downloaded through CRAN.
Listed below are program versions used for analysis.
- Python: 3.5.2
- Anaconda: 4.1.6
- Numpy: 1.11.0
- Scipy: 0.17.1
- Scikit: 0.17.1
- Pandas: 0.17.0
- R: 3.2.4
- ggplot2: 2.2.1
- reshape2: 1.4.2
- ply: 1.8.4
- dplyr: 0.5.0
- userfriendlyscience: 0.5-2
- RVAideMemoire: 0.9-62
- grid 3.2.4
- gridExtra 2.2.1
- for handling common file types
- needs to be on the python path or in the same directory as other scripts for all/most other scripts to run correctly
- extra utility functions
- needs to be on the python path or in the same directory as other scripts for all/most other scripts to run correctly
Combine multiple allC files at basepair level into one allC file
Usage: python combine_allc_pe.py [-f] [-p=num_proc] [-o=out_id] [-c=chrm_list | -cf=fasta_index]
<allc_path> <sample_name> [sample_name]*
Required:
allc_path path to allC files
sample_name name of sample; used to find allC files
when "-f" flag set, file with sample names listed one per line
Optional:
-f sample names are in file
-p=num_proc number of processors to use [default 1]
-o=out_id output file identifier [default "combined"]
-c=chrm_list comma-separated list of chrms to use
-cf=fasta_index fasta index file with chrms to use
Creates new allc files for all input samples that only includes information about positions which have at least minCov reads for each sample
expects all chromosomes in one allC file
Usage: python filter_allc_coverage.py [-v=min_cov] <allc_path> <sample1> [sampleN]*
Required:
allc_path path to allC files
sampleN name of sample; used to find allC files
Optional:
-v=min_cov min coverage for positions to include [default 3]
-p=num_proc number of processors to use [default 1]
Creates pseudo-allC file where all positions are unmethylated
Output file has same name as input file with "-unmethylated" appended
Usage: python unmethylate_allc_pe.py [-f] [-p=num_proc] [-v=NA] <allc_file> [allc_file]*
Required:
allc_file allC file to unmethylated
when "-f" set, file with list of allC files
Optional:
-f allC files names listed in the file
-p=num_proc number of processors to use [default 1]
-v=coverage coverage for each position [default as-is in input]
Used for between-generation computations of methylation
expects all chromosomes in one allC file and when minCov is set, minCov is output from filter_allc_coverage.py
Usage: python dmr_gen_counts.py [-o=outID] [-m=methType] [-p=numProc] [-v=min_cov]
<dmrFile> <allcPath> <sample1> <sample2> [sampleN]*
Required:
dmrFile tab-delimited file (BED format) with DMRs to investigate
allcPath Path to allC files; all chrms together for each sample
sample sample names as part of the allC file
Optional:
-o=outID identifier for output file [default "out"]
-m=methType methylation type [default C]
-p=numProc num. of processors to use [default 1]
-v=minCov min. coverage used as part of allC file name [default None]
Identifies significant methylation changes between generations using Fisher's exact test and minimum change in methylation
Input file is output from compare_dmrs_gens_pe.py
Usage: python dmr_gen_switches.py [-wm] [-n=num_c_thresh] [-m=meth_thresh] [-d=length_thresh]
[-f=fdr] [-o=outID] <in_file>
Required:
in_file tab-delimited file of DMRs and read counts
Optional:
-wm methylation threshold is for raw methyl difference
not percent difference
-n=num_c_thresh min number of cytosines in region to be considered for
analysis [default 10]
-d=lenth_thresh min length of dmr in bp [default 40]
-m=meth_thresh min methylation change btwn generations to be considered a
switch [default 0.3]
-f=fdr FDR value for significant switches [default 0.05]
-o=out_id identifier for output files [default uses input file name]
Converts the switches output of dmr_gen_switches.py to BED file
Usage: python dmr_file_to_bed.py [-v=score_thresh] [-p=name_prefix] [-o=outID] <in_file>
Required:
in_file input file of DMRs
Optional:
-v=score_thresh min score to include in output [default -1, no threshold]
-p=name_prefix prefix for naming features [default None]
-o=outID identifier for output file [default uses input file name]
Computes weighted methylation over regions
Usage: python dmr_counts_pe.py [-o=outID] [-m=methTypes] [-p=numProc] [-v=minCov] <dmrFile>
<allcPath> <sample1> <sample2> [sampleN]*
Required:
dmrFile tab-delimited file (BED format) with DMRs to investigate
allcPath Path to allC files; all chrms together for each sample
sample sample names as part of the allC file
Optional:
-o=outID identifier for output file [default "out"]
-m=methType methylation context to include [default C]
-p=numProc number of processors [default 1]
-v=minCov min. coverage used as part of allC file name [default None]
get individual positions that differ based on binomial test
Uses allC files specific to each chromosome
Usage: python find_all_mpos_dif_pe.py [-v=min_cov] [-c=chrm_list] [-o=out_id] [-p=num_proc]
[-m=meth_types] <allc_path> <sample1_name> <sample2_name>
Required:
allc_path path to allc files
sample_name names of samples to compare
Optional
-v=min_cov min coverage to include a position [default 3]
-o=out_id string for output file name [default "out"]
-c=chrm_list comma-separated list of chrms [default arabidopsis]
-p=num_proc num processors to use [default 1]
-m=meth_types comma-separated list of "CG", "CHG", and/or "CHH" [default all]
Filter a list of positions by gene-body methylation and/or CDS
Usage: python filter_pos_gene_gbm.py [-cds] [-v] <pos_file> <gbm_file> <gff_file>
Required:
pos_file position file, tab-delimited BED format, to be filtered
gbm_file file with list of gbM genes, one gene per line
use "none" or "na" to use all genes
gff_file GFF formatted file with genes
Optional:
-cds use CDS annotation not gene
-v include coordinates opposite of what is specified
Compute weighted methylation at each position specified, eliminating positions not covered by minCov reads in all samples
Position list is from find_all_mpos_dif_pe.py or filter_pos_gene_gbm.py
Usage: python weighted_meth_by_pos_pe.py [-o=out_id] [-v=min_cov] [-p=num_proc] <pos_list>
<allc_path> <sample_name> [sample_name]*
Required:
pos_list tab-delimited list with chrm and bp position
allc_path path to allc files
sample_name names of samples to include
Optional:
-v=min_cov min coverage to include a position [default 3]
-o=out_id string for output file name [default "out"]
-p=num_proc num processors to use [default 1]
Utility script used by epigenotyping_pe_combbin_fb-vit_cent.py; includes code for forward-backward decoding and Viterbi decoding
Utility script used by epigenotyping_pe_combbin_fb-vit_cent.py; computes the transition matrix
Major script which generates epigenotype map of samples based on mother and father methylomes
Input file is output of weighted_meth_by_pos_pe.py
Usage: python epigenotyping_pe_v1.7.3.py [-q] [-n-mpv] [-t-out] [-g=generation]
[-c=bin_thresh] [-d=decoding_type] [-p=num_proc] [-o=out_id] [-m=mother_
samples][-f=father_samples] [-b=bin_size] [-t=centromere] <input_file>
Requried:
input_file tab-delimited file of of weighted methylation by position for samples
Optional:
-q quiet; do not print progress
-h print help and exit
-n-mpv do not check for systematic mid-parent bias
-t-out write transition matrix to file
-g=generation generation of self-crossing; used to determine classification
probabilities; use 0 for uniform weight [default 2]
-d=decode_type decoding type to use (capitlization ignored) [default B]
Viterbi="v" or "viterbi"
Forward-Backward="forwardbackward", "f" or "fb"
All (FB and Vit independently)="all" or "a"
Both (FB then Vit)="both" or "b"
Off="false", "none", or "n"
-o=out_id identifier for output file [default "out" or variation of
input file name]
-p=num_proc number of processors [default 1
-c=bin_thresh minimum number of features per bin to be classified
groups bins to reach this number [default 3
-m=mother_label comma-separated sample name(s) of mother
[default mother]
-f=father_label comma-separated sample name(s) of father
[default father]
-b=bin_size size of bins in bp [default 100kbp]
-t=centromere centromere coordinates as "start,end"; can include multipe
centromeres as "start1,end1,start2,end2..." [default None]
Compute various accuracy scores comparing the assigned epigenotype and predicted epigenotype
Input file is created from R script, columns bin, sample, prediction, test, expected
Usage: python simulation_accuracy.py [-q] [-o=out_id] <input_file>
Required:
input_file csv file with expected and predicted epigenotype
Optional:
-o=out_id output identifier
-q quiet; don't print progress
Identify crossovers from an epigenotype map
Input file is output of epigenotyping_pe_combbin.py
Usage: python find_crossovers.py [-c=prediction_column] [-o=out_id] <input_file>
Required:
input_file tab delimited file with samples epigenotype per bin
Optional:
-o=out_id output identifier [default variation of
input file name
-c=prediction_column label of column to use as final epigenotype
[default "vit.prediction"]
Decodes and combines input pileup files into easier to read format
Assumes all input pileup files contain the same positions
In the output file, [A,C,G,T] indicate forward-strand read and [a,c,g,t] indicate reverse-strand read
Usage: python decode_pileup_pe.py [-o=out_id] [-p=num_proc] <pileup_file> [pileup_file]*
Required:
pileup_file pileup file for a sample; output from samtools pileup
Optional:
-o=out_id identifier for output file [default "out"]
-p=num_proc number of processors [default 1]
Based on the unique nt at each position between mother and father samples, guesses the genotype of the samples. Positions not distinguishable between parents are eliminated.
Input file is output of decode_pileup_pe.py
Usage: python pileup_genotype_pe.py [-o=out_id] [-p=num_proc] [-m=mother_label] [-f=father_label] [-v-min_cov] <decoded_pileup_file>
Required
decode_pileup_file tab-delimited input file; output of decode_pileup_pe.py
Optional:
-o=out_id identifier for output file [default variation of
input file name]
-v=min_cov min number of reads needed to support genotype [default 1]
-p=num_proc number of processors [default 1]
-m=mother_label sample name of mother [default mother]
-f=father_label sample name of father [default father]