bedRMod format¶
The bedRMod format specification working group.
Version v1.8. July 2024
The bedRMod format specification¶
The bedRMod, previously known as the EU (epitranscriptome unified data exchange) format is similar to the ENCODE bedMethyl format (BED9+2), but includes a header. The name (4th column) must conform to the MODOMICS nomenclature for the modification short name, and the score (5th column) is a site-specific measure of confidence.
The bedRMod file is a tabulated count of base modifications from every sequencing read over each reference genomic position or modification site. It is a convenient representation of the information stored in the MM/ML tags in BAM alignment files.
Hint
Sci-ModoM requirements
A given dataset or bedRMod file can contain more than one modification, as reported in column 4 (MODOMICS short name), but this should be for the same RNA type. Currently, supported RNA types are WTS or whole transcriptome sequencing. A dataset or bedRMod file can only contain ONE RNA type, ONE technology, ONE organism (incl. cell type, tissue, or organ), and records from the same assembly. The best way to handle treatment and/or conditions is to have as many bedRMod files as required to describe the experimental protocol, and provide a meaningful title and metadata for each file.
Attention
Format specification for tRNA
Sci-ModoM currently does not handle tRNA annotation. We are actively working on this, and more information will be soon available…
The header section¶
Each line starts with a # and contains the header tag and its value, separated by a =, e.g. #fileformat=bedRModv1.8.
As bedRMod contains data for only one organism, only one reference to organism, assembly, annotation source, and annotation version is possible.
Tag |
Description |
|---|---|
fileformat |
Fileformat and version e.g. bedRModv1.8 |
organism |
NCBI taxid |
modification_type |
RNA |
assembly |
Assembly name using Ensembl nomenclature e.g. GRCh38 for human |
annotation_source |
Annotation source e.g. Ensembl |
annotation_version |
Annotation version e.g. 110 |
sequencing_platform |
Sequencing platform e.g. Illumina, ONT, etc. |
basecalling |
Basecalling model information where relevant |
bioinformatics_workflow |
Reference to bioinformatics workflow e.g. GitHub, or information relevant to score, coverage and/or frequency calculation |
experiment |
Information about experimental protocol, design, etc. or link to e.g. openBIS |
external_source |
Databank;ID of data e.g. GEO;GSEXXXXXX |
While all tags are required, sequencing_platform, basecalling, bioinformatics_workflow, experiment or external_source can be left
empty, although it is strongly advised to provide as much information as possible, e.g.
#fileformat=bedRModv1.8
#organism=9606
#modification_type=RNA
#assembly=GRCh38
#annotation_source=Ensembl
#annotation_version=110
#sequencing_platform=Illumina NovaSeq 6000
#basecalling=
#bioinformatics_workflow=https://github.com/...
#experiment=https://doi.org/...
#external_source=GEO;GSEXXXXXX,PMID;XXXXXXXX
Hint
Sci-ModoM requirements
For data upload, fileformat, organism, and assembly are validated. Only the latest specifications are used, and
only modification_type=RNA is allowed. The assembly tag entry must match exactly the assembly chosen from the dropdown menu
during upload, and it must follow the Ensembl nomenclature. Data does not have to be for a specific genome assembly, Sci-ModoM will
take care of lifting over all records to the most recent assembly for each organism. Additional user-defined tags or extra header
lines starting with # are ignored during upload.
The data section¶
The first nine columns generally follow the standard BED specification, but the name (4th column) must conform to the MODOMICS nomenclature for the modification short name, and the score (5th column) is a site-specific measure of confidence. The last two columns contain the coverage (10th column), or number of reads at this position, and the frequency (11th column), or the integer value capped at 100 representing the precentage of reads that are modified at this position. These columns must be separated by tabs and each row must have the same number of columns.
Field |
Description |
|---|---|
chrom |
Chromosome |
chromStart |
Modification start position (0-based numbering, included) |
chromEnd |
Modification end position (0-based numbering, excluded), e.g. chromStart + 1 |
name |
MODOMICS short name |
score |
Modification confidence (see below for details) |
strand |
Strand |
thickStart |
Typically same as chromStart |
thickEnd |
Typically same as chromEnd |
itemRgb |
Color e.g. for genome browser |
coverage |
Number or reads at this position |
frequency |
Precentage (0,100] of modified reads, or stoichiometry at this position |
Hint
Sci-ModoM requirements
To enable quantitative data comparison, the score (5th column) is defined as round(-log10(p value)), where p value is calculated from a statistical test. A value of 0 indicates missing data e.g. p values were not calculated. The coverage (10th column) can be 0, if the number of reads at this position is not available, but frequency (11th column) MUST always be present. Modification frequency or stoichiometry is a minimal requirement for quantitative data comparison.
Caution
bedRMod is essentially a BED-formatted file, it uses a 0-based, half-open coordinate system. If you use a 1-based index, all your modification sites will be off-by-one!
Warning
For data upload to Sci-ModoM, chromosomes (1st column) must be formatted following the Ensembl short format e.g. 1 and not chr1, or MT and not chrM. Only chromosomes are considered, records from contigs/scaffolds are discarded. The modification name (4th column) must match exactly the chosen modifications, according to the MODOMICS nomenclature for the modification short name. Rows with out-of-range values for score (5th column) or frequency (11th column) are discarded.
File upload will fail if there are too many skipped records, e.g. due to wrong chromosome formatting, too many contigs, out-of-range values, etc.
Additional columns¶
Users can add any number of additional columns to suit their needs (same as for BED), but these are ignored in Sci-ModoM. Note however that a bedRMod file with exactly 12 columns may be implicitely assumed to be a BED12 file by some software (bedtools, genome browsers, …), which can result in unexpected behaviour.
Notes¶
Unmodified bases¶
bedRMod is a format to store modification data (site-specific or not), hence unmodified bases should not be recorded. Context can be recorded using chromStart/End + thickStart/End, additional columns, etc.
Download¶
A PDF version of the latest specification can be downloaded here.