Hello, I have been using Transdecoder to predict ORFs within my de-novo Transcriptome and to translate nucleotide sequences to their relative peptide sequences. I have been relying on the output of this program (where it characterizes each predicted peptide as 'complete', '5-prime partial', '3-prime partial', or 'internal') to help determine which sequences I should be pursuing for cloning. It was my belief that I should be focusing on the "complete" peptides if I want to generate functional proteins, so I was disregarding all sequences that weren't listed as "complete". However, on closer inspection I realized that most of the 5-prime partial sequences contain a methionine relatively close to the predicted start of the ORF. The program simply passes these off as incomplete due to the first AA being something other than methionine.
I am wondering if I should be considering these sequences in my cloning experiments or if I should discard them; on the one hand these sequences are from a transcriptome so there is some expression presumably at play. On the other hand, I am unsure whether the first methionine in the sequence would function as a true "stop" if I were to synthetically cut off the amino acids before it. Your opinion would be appreciated. Thank you!
Hello,
we have done a lot of de novo transcriptome assemblies followed by protein prediction and functional annotation and we have found the same "issue". My explanation is that methionine are not only present as a start of the protein but also inside the sequence. Most probably the methionine you find at close distance from the the start of the predicted protein is just an internal methionine. Indeed, if the methionine you see was the actual start then you should have an UTR upstream of that with no in-frame translation. The fact that you see an in-frame translation before the methionine should point out that the transcript is incomplete at the 5'. You could check manually some of these proteins by BLASTing them and see if the hits are "longer" at the N-terminal.
In terms of cloning you can also consider to analyze the presence of functional domains in the "partial" proteins (for instance with interproscan) and decide to clone the sequence if it contains all the expected domains.
I hope this might be helpful,
have a nice day
Riccardo Aiese Cigliano
Bioinformatician at www.sequentiabiotech.com
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude