Skip to content

Latest commit

 

History

History
79 lines (60 loc) · 6.4 KB

definition.md

File metadata and controls

79 lines (60 loc) · 6.4 KB
Raw

Project Status: Moved to http://example.com – The project has been moved to a new location, and the version at that location should be considered authoritative. to https://github.com/miRTop/mirGFF3

As discussed here: #10 we'll try to define a GFF3 format for output of small RNA pipelines

authors: @lpantano @gurgese @ThomasDesvignes @mhalushka @mlhack @keilbeck @BastianFromm @ivlachos @TJU-CMC @sbb25

GFF3 definition: https://github.com/The-Sequence-Ontology/Specifications/blob/master/gff3.md

Goals

We'll focus on miRNA first, then export to other databases

  • define each column to be adapted to miRNA annotation
  • define column 9 to support information like count data for samples, isomiRs labeling etc

Note: Keep in mind this is for the output of a pipeline, so we know there will be bias toward methodology, but the idea is to put enough information to be able to re-analyze or filter sequences using information described here. As well, it would be a proxy for downstream analysis or packages.

Deadline: https://github.com/miRTop/incubator/milestone/2

Description

Please add description for each columnd/attribute

  • header:
    • database: ##source-ontology LINK TO DATABASE include version and link (mandatory)
    • commands used to generate the file. At least information about adapter removal, filtering, aligner, mirna tool. All of them starting like: ## CMD: . Can be multiple lines starting with this tag. (optional)
    • genome/database version used (maybe try to get from BAM file if GFF3 generated from it): ## REFERENCE:
    • sample names used in attribute:Expression: ## COLDATA: separated by commas (mandatory)
    • small RNA GFF version ## VERSION: 0.9
    • Filter tags meaning: See Filter attribute below. Different filter tags should be separated by , character. Example: ## FILTER: and example would be ## FILTER: PASS(is ok), REJECT(false positive), REJECT lowcount(rejected due to low count in data).
  • column1: seqID: precursor name
  • column2: source: databases (lower case) used for the annotation (miRBase, mirgeneDB,tRNA...etc): #13. With the version number after _ character: mirbase_21
  • column3: type: ref_miRNA, isomiR: #13 (SO:0002166 ref_miRNA and SO:0002167 isomiR)
  • column4/5: start/end: precursor start/end as indicated by alignment tool
  • column6: score (Optional): It can be the mapping score or any other score the tool wants to assign to the sequence.
  • column7: strand. In the case of mapping against precursor should be always +. It should accept mapping against the genome: +/- allowed.
  • column8: phase: (For features of type "CDS", the phase indicates where the feature begins with reference to the reading frame): Not relevant righ now. This can be: .
  • column9: attributes:
    • UID: unique ID based on sequence like mintmap has for tRNA: prefix-22-BZBZOS4Y1 (https://github.com/TJU-CMC-Org/MINTmap/tree/master/MINTplates). good way to use it as cross-mapper ID between different naming or future changes. The tool will implement this, so an API can be used to fill this field. Currently supported by mirtop code.
    • Read: read name
    • Name: mature name
    • Parent: hairpin precursor name
    • Variant: (categorical types - adapted from isomiR-SEA)
      • iso_5p:+/-N. + indicates extra nucleotides not in the reference miRNA. - indicates removed nucleotides not in the sequence. N the number of nucleotides of difference. For instance, if the sequence starts 2 nts after the reference miRNA, the label will be: iso_5p:-2, but if it starts before, the label will be iso_5p:+2.
      • iso_3p:+/-N. Same explanation applied.
      • iso_add:+N. Same explanation applied.
      • iso_snp_seed: when affected nucleotides are betweem [2-7].
      • iso_snp_central_offset: when affected nucleotides is at position [8].
      • iso_snp_central: when affected nucleotides are betweem [9-12].
      • iso_snp_central_supp: when affected nucleotides are betweem [13-17].
      • iso_snp: anything else.
    • Changes (optional): similar to previous one but indicating the nucleotides being changed.
      • additions are in capital case
      • deletions are in lower case
      • exaple: Changes iso_5p:0,iso_3p:TT,iso_add:GTC,iso_snp:0 where Variant iso_add:+3,iso_3p:-2.
    • Cigar: CIGAR string as indicated here. It is the standard CIGAR for aligners. With the restriction that M means exact match always. That's a difference with some aligners where M includes mismatches. In this case, if there is a mismatch, then it should be output like: 11MA7M to indicates there is a mismatch at position 12, where A is the reference nucleotide.
    • Hits: number of hits in the database.
    • Alias (Optional): get names from miRBase/miRgeneDB or other database separated by ,
    • Genomic (Optional): positions on the genome in the following format: chr:start-end,chr:start-end
    • Expression: raw counts separated by ,. It should be in the same order than colData in the header.
    • Filter: PASS or REJECT (this allow to keep all the data and select the one you really want to conside as valid features). PASS can have subclases: PASS:te: meaning the sequence pass but the tools consider variants showed here are not trusted. REJECT can go with any short word explaining why it was rejected: REJECT:lowcounts. In this case the sequence will be skipped for data mining of the file when quering counts or summarize miRNA expression.
    • Seed_fam (Optional): in the format of 2-8 nts and reference miRNA sharing the seed. Usefull to go for pre-computed target predictions: ATGCTGT:mir34a_5p

API

Developing an API to check the format and help with the following feature will help to get people using it. Features should have:

  • command line and API
  • from BAM file (precursor/genome) and miRNA annotation files create the GFF3 file
  • parse GFF3 to filter, get stats, or convert to tabular format. something similar to samtools/vcftools, etc...
  • get fasta sequence from GFF3 attribute and genome FASTA file
  • (Add more)

I think python is a good way, installing via pipy or conda should get people using it pretty quickly.

Repo: https://github.com/miRTop/mirtop