Do you have experience in creating a protein database for a new proteome? We have a project with a collaborator who has some transcriptomic (RNA-seq) information but there is no (or very little) information available at the protein level i.e. there is not a database that can be downloaded from Uniprot.
We are able to generate .raw MS and MS/MS data for a protein lysate of the organism with ease, but then how does one go about identifying the peptides?
Does anyone have knowledge of de novo sequencing for complex proteomics?
Kind regards, Peter.
Hi Peter,
de novo detection of peptides from MS data is not as sensitive as Database search methods. If you have the transcriptome already sequenced, you can use it for peptide discovery by creating a protein database by theoretically translating (Forward three frame translation) the de novo assembled transcripts. For de novo assembly of transcritome you can use Oases or Trinity packages.
If the translated transcrptome database is huge, you can run an orf prediction using estscan to get potential polypeptide sequences to reduce the size of database.
Additionally, you can search MS data against a proteome database from closely related species.
Best,
Dhirendra
In addition to the above good suggestion: A brute force approach. if you download the entire NCBI and search a small mgf like file (depending on your server it can take only few hours), then you look at which species it is most likely close to (and generate a small database with the most common species found). If you already know that, then you can make a merged fasta database of those close species...Far from being perfect, it can help to ID an important fraction of your peptides.
T
Thank you Dhirendra & Thierry - both good suggestions.
After discussions with the collaborators, a FASTA database for theoretical proteins has been generated for the organism.
Many thanks, Peter.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics
- Bioinformatic Tools