GeenuFF

Specific changes vs gff

features

types

Feature types differ somewhat from a gff, “geenuff_” has been appended to names to reduce confusion where the meaning is not identical.

core types:

• geenuff_transcript: range of pre-mRNA / unspliced transcript
• geenuff_cds: range between the start and stop codon (ignoring introns)
• geenuff_intron / trans intron: range between a donor and acceptor splice site

error types:

These do not exist in a gff, but are used in geenuff to denote things that might be ambiguous or unknown about a gene model.

Currently errors are assigned when any obvious gene model inconsistency is encountered during gff parsing. Most error types are extended from the end of a known feature half way to the next gene model, while some internal errors (e.g. too_short_intron) can be assigned more precisesly. If gff format were not used as an intermediary and gene annotation was performed and stored directly in a geenuff structured database, all the errors could be assigned more precise ranges for any ambiguity.

Types are:

• missing_utr_5p
• missing_utr_3p
• empty_super_locus
• missing_start_codon
• missing_stop_codon
• wrong_starting_phase
• mismatched_ending_phase
• overlapping_exons
• too_short_intron
• super_loci_overlap_error

start_is_biological_start and end_is_biological_end:

When True, these attributes mean the start and end attributes of a feature correspond to a meaningful biological transition.

• start: start of a region, inclusive
  • geenuff_transcript –> transcription start site (1st transcribed bp)
  • geenuff_cds –> start codon, the A of the ATG
  • geenuf_intron –> donor splice site, first bp of intron
• end: end of a region, exclusive (i.e. start of one there after)
  • geenuff_transcript –> 1 after transcription termination site (1st non-transcribed bp)
  • geenuf_cds –> 1 after stop codon, first non-coding bp, e.g. the N in TGAN
  • geenuff_intron –> 1 after acceptor splice site (1st bp that is part of final transcript)

When False, these attributes mean the start and end attributes of a feature either do not, or it is not known if they correspond to a biological transition, yet the region they delineate is confidently of the given type.

For instance, if the parser finds a gene model in a gff where the start of the first exon and the start of the first CDS (+ strand) have the same position (the A in ATG), then it is apparent that we are missing the 5’ UTR, so for the geenuff_transcript feature the start_is_biological_start will be set to False, and an error mask will be added upstream of the CDS. We are still confident that all of the CDS must occur within the transcript, we know the start codon is part of the transcript region, but we mark that the start point itself is probably wrong, and mask the upstream range as it’s unclear what part of this is intergenic and which part UTR.

feature start/end/at numbering

Features have start and end coordinates that delineate a range.

The positioning of these features is in keeping with the common coordinate system: count from 0, start inclusive, end exclusive. So, the “geenuff_cds, start”, is at the A, of the ATG, AKA the first coding base pair; while in contrast, the “geenuff_cds, end” is after the stop-codon, AKA, the first non-coding bp.

reverse complement

Importantly, the coding-start should always point to the first A, of ATG, regardless of strand. This means the numeric coordinates have to change and unfortunately while one could take the sequence [1, 4) on the + strand, and directly use 1 and 4 as python coordinates and get the sequence; the same is not going to work on the minus strand. Instead:

 0  1  2  3  4  5
.N [A .T .G )N .N
 |  |  |  |  |  |
 N. T. A. C. N. N.

To get the reverse complement of this on the minus strand, we set the inclusive start to 3, and exclusive end to 0. Note this is now off by one from the python coordinates

 0  1  2  3  4  5
.N [A .T .G )N .N
 |  |  |  |  |  |
 N( T. A. C] N. N.

differences vs gff

Cheat sheet for how the Features compare to the gff (in particular any discrepancy between the closest coordinate in the gff, and the now standardized, consistent coordinate).

First and last for gff are reported as they are typically in gff (coordinate sorted), so reverse to the interpretation when on the - strand.

Plus strand (+)

Common Name	GFF	GFF start	GFF end	geenuff type	bearing	position
TSS, Transcription start site	start 1st exon	x		geenuff_transcript	start	x - 1
TTS, Transcription termination site	end last exon		x	geenuff_transcript	end	x
1st bp of start codon	start 1st CDS	x		geenuff_cds	start	x - 1
coding end	end last CDS		x	geenuff_cds	end	x
donor splice site (5’ of intron)	end non-last exon		x	geenuff_intron	start	x
acceptor splice site (3’ of intron)	start 2nd+ exon	x		geenuff_intron	end	x - 1

Minus strand (-)

Common Name	GFF	GFF start	GFF end	genuff type	bearing	position
TSS, Transcription start site	end last exon		x	geenuff_transcript	start	x - 1
TTS, Transcription termination site	start 1st exon	x		geenuff_transcript	end	x - 2
1st bp of start codon	end last CDS		x	geenuff_cds	start	x - 1
coding end	start 1st CDS	x		geenuff_cds	end	x - 2
donor splice site (5’ of intron)	start 2nd+ exon	x		geenuff_intron	start	x - 2
acceptor splice site (3’ of intron)	end non-last exon		x	geenuff_intron	end	x - 1

This site is open source. Improve this page.