DMRs

Detection of regional differences in methylation between groups.


BAT_DMRcalling

BAT_DMRcalling offers two modi: you either can use it, to call DMRs with metilene and filter these DMRs on the fly by a set of parameters independent of the metilene parameters, or you feed it with already called DMRs and use the post-calling filter option only. The output are the metilene output, without any post-calling filtering, a filtered bedGraph file containing the methylation rate differences per DMR, and a BED file with an unique identifier per DMR and a hypo-/hyper-methylation label.

In addition, basic statistics of the filtered DMRs are visualized including a histogram of the methylation differences, length distributions (in Cs and nucleotides), scatterplots of mean methylation rates of both groups and methylation difference vs. q-value for each DMR.

Basic usage

BAT_DMRcalling  -q <file> -o <prefix>

Output files

File Description
metilene.out BedGraph-like file containing the output of the DMR caller metilene.
metilene_filter.out BedGraph-like file containing the filtered output of the DMR caller metilene.
metilene_filter.bedgraph BedGraph file containing the difference of the mean methylation rate per group for each DMR passing the filtering criteria.
metilene_filter.bed BED-like file containing a unique identifier and the direction (hypo- or hyper-methylated) for each DMR passing the filter criteria.
metilene_filter.pdf PDF file with DMR statistics plots.

Input/Output options

Option Description
-q Input file for DMR calling with metilene (in case of mode 1 with option -F 1, default) or file produced by metilene for filtering (in case of mode 2 with option -F 2).
-o Prefix of output files.

Algorithm/Filtering options

Option Description
-F Select mode of operation. Mode 1: Call DMRs using metilene with default settings and group names (i.e., group1 and group2), followed by the filtering of the called DMRs using the filtering options below (default). Mode 2: Filter previously called DMRs given by input file using the filtering options below (default: 1).
-p Only report DMRs with adjusted p-value (i.e., q-value) below this threshold (post-processing, default: 0.05).
-c Minimum number of Cs (post-processing, default:10).
-d Minimum difference of mean methylation rates per group (post-processing, default: 0.1).
-l Minimum length in nucleotides (post-processing, default: 0).
-a Name of group 1 (default: g1).
-b Name of group 2 (default: g2).
-z Additional parameter for metilene (e.g., number of threads and distance between Cs with "-t <num> -M <num>").

External tools

Option Description
--metilene Path to metilene executable. Required if metilene executable is not in PATH. For installation, manual or problems please go to the metilene website.
-R Path to R executable. Required if R executable is not in PATH. For installation, manual or problems please go to the R website.
(top)

BAT_correlating

BAT_correlating can be used to inspect whether the methylation level over some regions (e.g., DMRs, transcription factor binding sites, CpG islands) correlate significantly with the expression of the associated genes. It simply requires a BED file to associate methylation regions with gene identifiers, a file to associate samples with the two groups, and files with expression values (for each gene identifier) as well as position-wise methylation rates (as bigWig files) for each sample.

For each region-gene pair, BAT_correlating will report a statistics with the linear and non-linear (Sperman's) correlation coefficient between expression and methylation including a p-value of Sperman's rank correlation test as well as a correlation plot showing the mean methylation rate of the region and the expression of the associated gene for each sample (as scatter and boxplot).

Basic usage

BAT_correlating -b <file> -e <file> -m <file> -i <group1>,<group2>,<group3> -g <file> -o <file>

Output files

File Description
prefix.statistics Text file containing gene identifier, methylation region, adjusted R-squared of the linear model, rho value (i.e., Spearman's rank correlation coefficient) of the non-linear model including a p-value of Spearman's rank correlation test for each genomic region to gene pair.
prefix_region_gene.pdf PDF file with the correlation plot showing boxplots of mean methylation and expression values per group, and a scatterplot of mean methylation and expression value of all samples (coloured by group) including regression line and information of linear model and non-parameteric model (Spearman's rank correlation).

Input/Output options

Option Description
-b BED file containing coordinates of region (for overlapping with methylation bigWig file) and name of associated gene identifier (to connect to expression file). Format: chr <tab> start <tab> end <tab> identifier. Identifier need to match gene identifiers of the expression files (option -e).
-e File containing a list of expression files (one per sample). The content of each file must follow this format identifier <tab> something <tab> expression_value.
-m File containing a list of methylation bigWig files (one per sample).
-g File containing the association of sample identifiers to groups. Format: sample <tab> group. The list of samples in the file must be ordered analogously as in the list of files given with options -e and -m.
-i Comma-separated list of group 1 and group2 as well as an optional additional third group for plotting. Groups need to match the group names in the file provide with option -g.
-o Path/prefix for output (path for correlation plots and path/prefix.txt for statistics).

Additional options

Option Description
--min Minimum number of length (in nucleotides) of regions in the BED file to be included in the correlation calculation (default: 5).
--prov Indicates that files given with options -e and -m already contain methylation and expression values, respectively. Format of file given by option -e: sample <tab> region <tab> identifer <tab> methylation_rate. Format of file given by option -m: sample <tab> identifer <tab> expression_value. Hence, only options -e, -m, and -b are required.

External tools

Option Description
--bw Path to UCSCtools' bigWigAverageOverBed executable. Required if bigWigAverageOverBed executable is not in PATH. For installation, manual or problems please go to the UCSCtools website.
-R Path to R executable. Required if R executable is not in PATH. For installation, manual or problems please go to the R website.
--latex Path to latex executable. Required if latex executable is not in PATH.
--dvipdf Path to dvipdf executable. Required if dvipdf executable is not in PATH.

(top)