Skip to content

Latest commit

 

History

History
53 lines (36 loc) · 3.47 KB

methods.md

File metadata and controls

53 lines (36 loc) · 3.47 KB
Raw

Methods

Samples of mentioned files

Gene2Refseq

NCBI_ID Transcript_ID Gene_Symbol
114108587 NM_001366559.1 ATF7-NPFF
114022705 - LOC114022705
114022708 - LOC114022708
114108587 NM_001366560.1 ATF7-NPFF
114108587 NR_159377.1 ATF7-NPFF

odb*_genes.tab OrthoDB README

og_id tax_id prot_name synonyms() description
9606_0:000000 9606_0 YP_003024037.1 MT-ND6;ND6 NADH-ubiquinone oxidoreductase chain 6
9606_0:000001 9606_0 YP_003024033.1 MT-ND3;ND3 NADH-ubiquinone oxidoreductase chain 3
9606_0:000002 9606_0 YP_003024027.1 MT-ND2;ND2 NADH-ubiquinone oxidoreductase chain 2
9606_0:000003 9606_0 YP_003024031.1 ATP6;MT-ATP6 ATP synthase subunit a
9606_0:000004 9606_0 YP_003024032.1 COX3;MT-CO3 Cytochrome c oxidase subunit 3

Essential dictionaries construction

  • refseq dictionaries: Parse the gene2refseq.txt to construct two dicts:

    • gene_name_to_ncbigid : {geneSymbol:NCBI_ID}
    • ncbi_to_refseq : map NCBI IDs to RefSeq transcript ID(s) , it's a one:many relationship.

  • ODB dictionaries:

    • odb_ogid_desc : Mapping Orthologus Gene Unique ID to Gene Description by parsing the odb10v0_genes.tab file
    • odb_ncbi_ogid : Mapping NCBI ID to Orthologus Gene Unique ID as following:
      1. parsing the odb10v0_genes.tab file to get Gene Symbol and Orthologus Gene Unique ID
      2. Quary the gene_name_to_ncbigid to get the NCBI ID, if not found
      3. deep search option could be selected (Slows down the process as it make online query)

Mapping Heuristics

Let's short-name the files, "odb10v0_genes.tab" > odb_genes & {species}_rna.fna.gz > refseq_fa

First, check the relation between the Orthologus Gene Unique ID and its corresponding Transcript IDs in the refseq_fa file via NCBI IDs

The relation could be categorized into one of 3 types:

Case 1. [1:1] Relationship If the NCBI ID occurred in a single odb_genes record, and it has just one corresponding Transcript ID in the refseq_fa

Case 2. [1:M] Relationship If the NCBI ID occurred in single odb_genes record, and it has multiple corresponding Transcript IDs in the refseq_fa. We try to map the Orthologus Gene Unique ID to the exact transcript of the corresponding NCBI ID using the description in the odb_genes file. But if we could not, the Orthologus Gene Unique ID will be mapped to all the transcripts of this gene.

Case 3. [M:M] Relationship If the NCBI ID occurred in multiple odb_genes records, and it has multiple corresponding Transcript IDs in the refseq_fa. Again we try to map the Orthologus Gene Unique ID to the exact transcript of the corresponding NCBI ID using the description in the odb_genes file. But if we still ending up with many Orthologus Gene Unique ID mapping to many transcripts, we drop this realtionship completely to avoid ambiguity

Note: The description in the odb_genes file uses the word "isoform" while the refseq_fa uses the word "transcript variant" to describe the transcripts so we consider this during the matching