download fasta file from uniprot

Question

[pinheirofabiano]

Hi,

I need to download a fasta file from uniprot, containing a list of the protein sequences of every secretory protein described to date with 200 or less amino acids.

Can somebody help me to filter these proteins and generate a list with IDs and sequences?

I'm trying to filter the proteins on the website, but I'm not finding how to do it. Every known secretory protein, from all species.

Many thanks,
Fabiano

[GenoMax]

How about: https://www.uniprot.org/uniprotkb?dir=descend&facets=length:[1%20TO%20200]&query=secretory&sort=length

If you are happy with this result, click on "Download" and make sure to "choose all" to get the fasta file.

[GenoMax]

Subset from Bacteria can be obtained using this --> LINK

Subtract the result from above to get non-bacterial sequences.

Answer 1

[Elisabeth Gasteiger]

If you are looking for secretory proteins in the sense of https://en.wikipedia.org/wiki/Secretory_protein , i.e. "A secretory protein is any protein, [...] which is secreted by a cell", I would recommend to search for entries which are annotated with the "Subcellular location" term "Secreted", https://www.uniprot.org/locations/SL-0243 rather than performing a full text search with the word "secretory". The latter will look for the word "secretory" anywhere in a record, including free text comments or publication titles (where it may well state that a protein is not secretory).

The entries are https://www.uniprot.org/uniprotkb?query=(cc_scl_term:SL-0243)

and with the additional constraint on sequence length: https://www.uniprot.org/uniprotkb?dir=ascend&query=(cc_scl_term:SL-0243)%20AND%20(length:[1%20TO%20200])&sort=length

You can obtain these queries by using UniProt's advanced search (click on "advanced") to the right of the query box.

Note that not all proteins matching the word "secretory" in the protein name are actually secreted. It is important to know what you are looking for exactly.

Please do not hesitate to contact the UniProt helpdesk if you have any questions about UniProt. If you wish to develop a tool that predicts certain properties, it is extremely important to get your training set right and in that case exclude unreviewed entries.