$ npm install --save generic-sequence-panel
This package provides two React components: GenericSeqPanel and GenericGeneSeqPanel which have similar arguments. The difference between these two panels is this: GenericSeqPanel provides ONLY the highlighted FASTA given a gene, transcript and mode. GenericGeneSeqPanel does a little more: you provide a gene and it gets a list of transcripts of that gene and provides a dropdown menu for both the transcript and the mode.
The React components create a div containing highlighted FASTA DNA sequence. It uses as its input materials an NCList (i.e., a JBrowse 1 GFF3 formatted dataset) and a bgzip, faidx indexed FASTA file. It addition to those data sources, the component must also be supplied these items about the location:
- the name of the reference sequence (eg, "Chr1")
- the start and end coordinates (technically in interbase coordinates, but since the lookup is going to be an overlaps-type query, the details don't really matter if base versus interbase coordinates are used)
- the name of the gene and transcript (only need for GenericSeqPanel) for which the highlighted fasta is required
- the "mode" of highlighting required (more on that below, and also only required by GenericSeqPanel)
<GenericSeqPanel
refseq="X"
start={13201770}
end={13216729}
gene="WBGene00006749"
transcript="R12H7.1a.2"
mode="protein"
nclistbaseurl="https://s3.amazonaws.com/agrjbrowse/MOD-jbrowses/WormBase/WS287/c_elegans_PRJNA13758/"
urltemplate="tracks/Curated_Genes/{refseq}/trackData.jsonz"
fastaurl="https://s3.amazonaws.com/wormbase-modencode/fasta/current/c_elegans.PRJNA13758.WS284.genomic.fa.gz"
/>
and
<GenericGeneSeqPanel
refseq="X"
start={13201770}
end={13216729}
gene="WBGene00006749"
nclistbaseurl="https://s3.amazonaws.com/agrjbrowse/MOD-jbrowses/WormBase/WS287/c_elegans_PRJNA13758/"
urltemplate="tracks/Curated_Genes/{refseq}/trackData.jsonz"
fastaurl="https://s3.amazonaws.com/wormbase-modencode/fasta/current/c_elegans.PRJNA13758.WS284.genomic.fa.gz"
/>
Several items here a self explanatory: refseq, start, and end are location information. The props gene and transcript are the names of the features in the NClist data set. The remaining items are described here:
- mode - this is one of several keys that dictate what the output looks like.
The options are:
- genomic - The stretch of sequence from start to end with no special highlighting
- genomic_sequence_updown - The stretch of sequence from start to end with 500 base pairs of padding on both ends
- cds - The coding sequence of the mRNA that is the result of in silico splicing
- cdna - The CDS with UTRs added
- protein - The amino acid sequence that results from the cds in silico transcription
- gene - The genomic sequence from start to end with portions that are UTR and coding highlighted
- gene_collapsed_intron - same as gene, but the introns are compressed to 10 base pairs at the splice junction and the remainder replaced with ellipses
- gene_updownstream - same as gene but with 500 bp of up and down stream sequence added
- gene_updownstream_collapsed_intron - same as gene_collapsed_intron but with up and downstream padding added
- nclistbaseurl - the base url for the NClist adapter (basically, it's the part before "tracks" in the url)
- urltemplate - the rest of the url that the NClis adapter uses. Typically, this
will have
{refseq}in it that will be interpolated with the refseq info - fastaurl - the url to the fasta file. The location of the .fai and .gzi files will be assumed from this url
This package also provides a helper function that is used internally to get the subfeature information for a given feature (gene). This function takes several arguments that are similar to the above components:
- nclistbaseurl: the base URL of the NCList file
- urltemplate: the URL template for the NCList file
- refseq: the reference sequence (name, like "chr1")
- start: the start of the range (in interbase coordinates)
- end: the end of the range (in interbase coordinates)
- gene: the name of the gene
and returns an array of JBrowse 2-style feature objects that correspond to the
gene's subfeatures (usually transcripts and their exons, UTRs and CDS regions).
In the console, those features look like the below, but note that they can also
be accessed via getter functions like transcriptArray[0].get('name') and
transcriptArray[i].get('start'), since the underlying JBrowse libraries handle
the mapping between the items in this array and the attributes that came from
GFF.
[[
1,
31097676, // start coordinate (interbase coords)
31169556, // end coordinate
-1, // strand
"ENSEMBL:ENST00000481143.2", // some sort of id
"DMD", // feature name
"rna97622", // some sort of id
"ENST00000481143.2", // another id
"SO:0000673", // Sequence ontology term ID
"X", // reference sequence name
"ENSEMBL", // GFF source?
[[
2, // subfeature depth?
31097676, // subfeature start
31098183, // subfeature end
-1, // subfeature strand
"e893041", // subfeature name
"X", // subfeature refrence sequence name
"ENSEMBL", // subfeature GFF source
"exon" // subfeature SO term name
],
[
2,
31169442,
31169556,
-1,
"e893042",
"X",
"ENSEMBL",
"exon"
]],
"ENSEMBL:ENST00000481143.2", // another id
"transcript" // SO term name
],
etc
An example of how this function is used can be found in the GenericGeneSeqPanel component:
https://github.com/scottcain/seqpanel/blob/main/lib/src/components/GenericGeneSeqPanel.tsx
This component makes use of three components developed by the JBrowse team:
- NCList from "@gmod/nclist" - This is what accesses the JBrowse 1 (NCList) data stores. It does an "overlaps" query of the NCList data set and returns all genes and their children (transcripts, exons, etc) that overlap. The code GenericSeqPanel then filters the feature set to the gene and transcript specified in the props of the component.
- BgzipIndexedFasta from "@gmod/indexedfasta" - This accesses the bgzipped, samtools faidx indexed fasta file. The locations of the .fai and .gzi files are found by just appending those extensions to the supplied fastaurl.
- SequencePanel from "@jbrowse/core/BaseFeatureWidget/SequencePanel" - This
component takes the feature data from NCList and the sequence data from
BgzipIndexedFasta and generates the highlighted sequence that is controlled by
the
modeprop described above.
This component does what I need it to do. Potential improviments include:
- Both NCList and BgzipIndexedFasta could be replaced with options to use other data accessing tools, like one for tabix-indexed GFF3 for feature data and twobit for reading .2bit sequence files.
- This component assumes all data are remote and accessed via URL. Options for local access could be added.
- It also assumes the "standard" genetic code. An option could be added to support others.
Pull requests are accepted.
In this screenshot, the actual output of the GenericGeneSeqPanel component is shown. The portion that is just the highlighted fasta is what is provided by GenericSeqPanel, and the rest of the UI (dropdown menus, buttons, and legend) are provided by GenericGeneSeqPanel.
A simple example implementation (from the app directory in this monorepo) is running at https://scottcain.github.io/seqpanel/
I would very much like to thank Colin Diesh, who wrote large chunks of the code in this repo while teaching me the ins and outs of the pieces of JBrowse code I wanted to use.
This package was written with funding from the NHGRI as part of the JBrowse project. If you use it in an academic project that you publish, please cite the most recent JBrowse paper, which will be linked from jbrowse.org.
MIT © Scott Cain