Chapter 15 Downloading sequences from UniProt by hand
By: Avril Coghlan.
Adapted, edited and expanded: Nathan Brouwer under the Creative Commons 3.0 Attribution License (CC BY 3.0).
15.1 Vocab
- RefSeq
- manual curation
- UniProt
- accession
15.2 Downloading Protein data from UniProt
In a previous vignette you learned how to retrieve sequences from the NCBI database. The NCBI database is a key database in bioinformatics because it contains essentially all DNA sequences ever sequenced.
As mentioned previously, a subsection of the NCBI database called RefSeq consists of high quality DNA and protein sequence data. Furthermore, the NCBI entries for the RefSeq sequences have been manually curated, which means that expert biologists employed by NCBI have added additional information to the NCBI entries for those sequences, such as details of scientific papers that describe the sequences.
Another extremely important manually curated database is UniProt www.uniprot.org, which focuses on protein sequences. UniProt aims to contains manually curated information on all known protein sequences. While many of the protein sequences in UniProt are also present in RefSeq, the amount and quality of manually curated information in UniProt is much higher than that in RefSeq.
For each protein in UniProt, the UniProt curators read all the scientific papers that they can find about that protein, and add information from those papers to the protein’s UniProt entry. For example, for a human protein, the UniProt entry for the protein usually includes information about the biological function of the protein, in what human tissues it is expressed, whether it interacts with other human proteins, and much more. All this information has been manually gathered by the UniProt curators from scientific papers, and the papers in which the found the information are always listed in the UniProt entry for the protein.
Just like NCBI, UniProt also assigns an accession to each sequence in the UniProt database. Although the same protein sequence may appear in both the NCBI database and the UniProt database, it will have different NCBI and UniProt accessions. However, there is usually a link on the NCBI entry for the protein sequence to the UniProt entry, and vice versa.
15.3 Viewing the UniProt webpage for a protein sequence
If you are given the UniProt accession for a protein, to find the UniProt entry for the protein, you first need to go the UniProt website, www.uniprot.org. At the top of the UniProt website, you will see a search box, and you can type the accession of the protein that you are looking for in this search box, and then click on the “Search” button to search for it.
For example, if you want to find the sequence for the chorismate lyase protein from Mycobacterium leprae (the bacterium which causes leprosy), which has UniProt accession Q9CD83, you would type just “Q9CD83” in the search box and press “Search”. The UniProt entry for UniProt accession Q9CD83 will then appear in your web browser.
Beside the heading “Organism” you can see the organism is given as Mycobacterium leprae. If you scroll down you’ll find a section Names and Taxonomy and beside the heading “Taxonomic lineage”, you can see “Bacteria - Actinobacteria - Actinobacteridae - Actinomycetales - Corynebacterineae - Mycobacteriaceae- Mycobacterium”.
This tells us that Mycobacterium is a species of bacteria, which belongs to a group of related bacteria called the Mycobacteriaceae, which itself belongs to a larger group of related bacteria called the Corynebacterineae, which itself belongs to an even larger group of related bacteria called the Actinomycetales, which itself belongs to the Actinobacteridae, which itself belongs to a huge group of bacteria called the Actinobacteria.
15.3.1 Protein function
Back up at the top under “organism” is says “Status”, which tells us the annotation score is 2 out of 5, that it is a “Protein inferred from homology”, which means what we know about it is derived from bioinformatics and computational tools, not lab work.
Beside the heading “Function”, it says that the function of this protein is that it “Removes the pyruvyl group from chorismate to provide 4-hydroxybenzoate (4HB)”. This tells us this protein is an enzyme (a protein that increases the rate of a specific biochemical reaction), and tells us what is the particular biochemical reaction that this enzyme is involved in. At the end of this info it says “By similarity”, which again indicates that what we know about this protein comes from bioinformatics, not lab work.
15.3.2 Protein sequence and size
Under Sequence we see that the sequence length is 210 amino acids long (210 letters long) and has a mass of 24,045 daltons. We can access the sequence as a FASTA file from here if we want and also carry out a BLAST search from a link on the right.
15.3.3 Other information
Further down the UniProt page for this protein, you will see a lot more information, as well as many links to webpages in other biological databases, such as NCBI. The huge amount of information about proteins in UniProt means that if you want to find out about a particular protein, the UniProt page for that protein is a great place to start.
15.4 Retrieving a UniProt protein sequence via the UniProt website
There are a couple different ways to retrieve the sequence. At the top of the page is a tab that say “Format” which brings you to a page with th FASTA file. You can copy and paste the sequence from here if you want. To save it as a file, go to the “File” menu of your web browser, choose “Save page as”, and save the file. Remember to give the file a sensible name (eg. “Q9CD83.fasta” for accession Q9CD83), and in a place that you will remember (eg. in the “My Documents” folder).
For example, you can retrieve the protein sequences for the chorismate lyase protein from Mycobacterium leprae (which has UniProt accession Q9CD83) and for the chorismate lyase protein from Mycobacterium ulcerans (UniProt accession A0PQ23), and save them as FASTA-format files (eg. “Q9CD83.fasta” and “A0PQ23.fasta”, as described above.
You can also put the UniProt information into an online Basket. If you do this for both Q9CD83 and A0PQ23 you can think click on Basket, select both entries, and carry out a pairwise alignment by clicking on Align.
Mycobacterium leprae is the bacterium which causes leprosy, while Mycobacterium ulcerans is a related bacterium which causes Buruli ulcer, both of which are classified by the WHO as neglected tropical diseases. The M. leprae and M. ulcerans chorismate lyase proteins are an example of a pair of homologous (related) proteins in two related species of bacteria.
If you downloaded the protein sequences for UniProt accessions Q9CD83 and A0PQ23 and saved them as FASTA-format files (eg. “Q9CD83.fasta” and “A0PQ23.fasta”), you could read them into R using the read.fasta() function in the SeqinR R package (as detailed in another Vignette) or a similar function from another package.
Note that the read.fasta() function normally expects that you have put your FASTA-format files in the the working directory of R. For convenience so you can explore these sequences they have been saved in a special folder in the compbio4all package and can be accessed like this for the Leprosy sequence
# load compbio4all
library(compbio4all)
# locate the file within the package using system.file()
file.1 <- system.file("./extdata/Q9CD83.fasta",package = "compbio4all")
# load seqinr
library("seqinr")
# load fasta
leprae <- read.fasta(file = file.1)
lepraeseq <- leprae[[1]]We can confirm
str(lepraeseq) # 'SeqFastadna' chr [1:210] "m" "t" "n" "r" "t" "l" "s" "r" "e" "e" "i" ...
# - attr(*, "name")= chr "sp|Q9CD83|PHBS_MYCLE"
# - attr(*, "Annot")= chr ">sp|Q9CD83|PHBS_MYCLE Chorismate pyruvate-lyase
# OS=Mycobacterium leprae (strain TN) OX=272631 GN=ML0133 PE=3 SV=1"For the other sequence
# locate the file within the package using system.file()
file.2 <- system.file("./extdata/A0PQ23.fasta", package = "compbio4all")
# load fasta
ulcerans <- read.fasta(file = file.2)
ulceransseq <- ulcerans[[1]]
str(ulceransseq)[[1]] # 'SeqFastadna' chr [1:212] "m" "l" "a" "v" "l" "p" "e" "k" "r" "e" "m" ...
# - attr(*, "name")= chr "tr|A0PQ23|A0PQ23_MYCUA"
# - attr(*, "Annot")= chr ">tr|A0PQ23|A0PQ23_MYCUA Chorismate pyruvate-lyase
# OS=Mycobacterium ulcerans (strain Agy99) OX=362242 GN=MUL_2003 PE=4 SV=1"