predicted protein and sequence

Question

[mdfardin374]

Hello,

Can somebody tell me what is predicted protein and how one can retrieve all the predicted proteins of a particular organism? Thank you

[gb]

Maybe this post helps C: Which software is suitable for protein prediction from whole eukaryotic genomes?

In combination with this website:

https://en.wikipedia.org/wiki/Gene_prediction

[Joe]

What do you suppose a "predicted protein" might be? Given that its name is an exact description of what it is.

Answer 1

[hugo.avila]

The predicted proteins are the output of a program that scans all possible coding sequences (CDS) (codon Start to Stop) in a genome and assigns a name and function based on matches with an experimental protein sequence database. To search, you need to find a program that fits your needs (prokaryotes or eukaryotes). If you just want to extract the CDSs from your genome, a simple biopython script can do that. But if you want to assign name and function (annotation), you need to use a program. I work with prokaryotes, the best program for me is Prokka.

[mdfardin374]

I want to retrieve predicted protein (the protein that has no experimental evidence) already present in the database but I do not know how to retrieve it

[GenoMax]

For a specific organism or in general?

[mdfardin374]

for specific organism

[GenoMax]

If a genome page is available for this organism at NCBI then follow these directions (this is a random example genome).

1. Find genome page of your organism at NCBI. I am choosing this one.
   
2. Find the protein link for the genome (should be up near top of the page).
   

3. Find predicted protein entries from the file.
   

   $ zmore GCF_000209225.1_ASM20922v1_protein.faa.gz | grep "predicted"

   XP_001623027.1 predicted protein [Nematostella vectensis] XP_001623028.1 predicted protein [Nematostella vectensis] XP_001623029.1 predicted protein [Nematostella vectensis] XP_001623030.1 predicted protein [Nematostella vectensis]

4. Extract the sequence for those ID's using faSomeRecords utility from Jim Kent at UCSC.

[hugo.avila]

From which database do you want to extract these protein sequences ? In most databases you can search for CDS titled "Hypothetical Protein". These CDS are predicted sequences that show no homology to a known sequence.