Read this first:

This repository was last updated for March-2017 UniprotKB release. Now, I have moved on to using BigBed files from UniProt, instead of Bed files used here. This recent work is available at repository uniprot_genomic.

UniProt in hg19 coordinates

UniProt provides human genome annotation data enabling mapping of amino acid annotations directly to reference genome coordinates, but they are available only in hg38 coordinates. See this publication for more info:

Nucleic Acids Res. 2017 Jan 4;45(D1):D158-D169. doi: 10.1093/nar/gkw1099. UniProt: the universal protein knowledgebase. The UniProt Consortium.

This repository converts and makes this data available in hg19 coordinates.

Files for download:

Besides conversion to hg19 coordinated, few changes are made here to suit our purposes, which is to identify if query amino acids have any UniProt annotation. See ‘Processing pipeline’ section for details.

• Restructured, hg19-converted Bed files. This is what you probably are interested in.

• Two merged files each containing selective sequence annotations of interest, as listed below.

  a. Merged file - Type 1 has following annotation types merged into a single file.

    1    Active site
    2    Binding site for any chemical group
    3    Calcium binding region
    4    Cross-link between proteins
    5    Disulfide bond
    6    Glycosylation-PTM
    7    Interesting site
    8    Lipidation-PTM
    9    Metal binding site
    10    Motif
    11    Nucleotide binding region
    12    Other PTM
    13    Signal peptide
    14    Transit peptide
    15    Zinc finger region

  b. Merged file - Type 0 has following annotation types merged into a single file.

    1    Active peptide
    2    Chain
    3    Coiled coil
    4    DNA binding domain
    5    Domain
    6    Intramembrane
    7    Natural variant
    8    Region of interest
    9    Repeated motifs or domains
    10    Topological domain
    11    Transmembrane region

Processing pipeline:

1. Use liftOver tool for conversion of hg38 to hg19 coordinates.
   Note: If you are interested in excecuting the script, download chain file and store it in settings_files directory.
   It is not provided here due to license concerns.
2. Fix formatting issues in resulting Bed files.

3. Reformat Bed files as follows:

   a. Replace score column (5th column), which is zero by default in UniProt provided data, with corresponding sequence annotation type as shown below.

    Original format by UniProt:
    >chr1    7970956    7970959    Q99497    0    +    7970956    7970959    255,102,102    1    3    0    .    Nucleophile. Pubmed:20304780, Pubmed:25416785
    Format we used here:
    >chr1    8031016    8031019    Q99497    Active site    +    8031016    8031019    255,102,102    1    3    0    .    Nucleophile. Pubmed:20304780, Pubmed:25416785

b. Restructure the rows in Bed files that have non-continuous amino acids as in example below.
```
Original format by UniProt (this line has coordinates for three, non-continuous amino acids):
>chr1    1633782    1633815    O75900    0    +    1633782    1633815    0,153,0    3    3,3,3    0,12,30    .    Zinc; catalytic.
Format we used here (one amino acid per row, if non-continuous):
>chr1    1569161    1569164    O75900    Metal binding site    +    1569161    1569164    0,153,0    1    3    0    .    Zinc; catalytic.
>chr1    1569173    1569176    O75900    Metal binding site    +    1569173    1569176    0,153,0    1    3    0    .    Zinc; catalytic.
>chr1    1569191    1569194    O75900    Metal binding site    +    1569191    1569194    0,153,0    1    3    0    .    Zinc; catalytic.
```

Resulting Bed files are what you probably need if you are looking for replacement for UniProt provided hg38 genome coordinates in hg19 format.

Further Restructuring:

We further merge sequence annotation types of our interest into two Bed files.

1. For annotation type ‘natural variant’, replace disease acronyms with their complete name.
2. Merge Bed files of interest (as customized in the settings file; based on values 0 and 1) based on sequence annotation types into two sets of merged files.

Download the resulting merged bed files:

a. Merged Bed file - Type 1

b. Merged Bed file - Type 0

Disclaimer

UniProt’s license applies for the genome coordinates data available in this repository. Thanks to UniProt for permitting us to distribute this data in hg19 format. Data is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Code in this repository is distributed under MIT license.