The RBS calculator has been used to predict protein expression in various bacteria or to design RBS sequences for a desired expression level. Do you have any experience about this tool?
I've used it several times before (different versions). It's great for finding and eliminating unwanted RBS from de novo gene synthesis projects. Hitting close to a particular absolute expression level is difficult, but in my experience usually if the Salis lab calculator predicts higher or lower relative expression, this will almost always be the case. Of course, my growth conditions are different than those used to refine this model, so I don't expect the same results at all, and the relevant papers don't claim perfect predictions (IIRC, measurements usually fall within 50-200% of the predicted value, but the model's been improved recently).
Sometimes for making very efficient RBS, I will directly look at RNA folding predictions for the 5' 100 bp or so and eliminating predicted secondary structure by mutating the 5' UTR and some codons.
The RBS Calculator can be used in either Reverse Engineering or Forward Engineering modes. In the Reverse Engineering mode, you enter the nucleotide sequence of an mRNA transcript and the RBS Calculator predicts the translation initiation rate from each of its start codons, either AUG, GUG, or UUG. In the Forward Engineering mode, you enter a protein coding sequence and a target translation initiation rate. The RBS calculator then generates a synthetic ribosome binding site sequence that will initiate your protein coding sequence at the selected translation initiation rate.
Forward Engineering Mode
Step 1: Copy and paste the first 50 nucleotides of a protein coding sequence into the Protein Coding Sequence box. The sequence must start with either AUG or GUG and any combination of ATCGU nucleotides is acceptable. Thymines (T) are automatically converted to uracils (U).
Step 2: Select the target translation initiation rate, which is gauged on a proportional scale from 0.1 to 100,000+. There is no lower or upper limit to the scale and translation rates below 0.1 or above 100,000 are feasible, depending on the protein coding sequence.
Step 3: Input the pre-sequence, which is any sequence that appears prior to the ribosome binding site, such as a restriction site. The sequence must be a part of the mRNA transcript.
Step 4: Click on the "Submit Job" button. The result will be stored in your "My Results" list and will be completed within 1 to 15 minutes, depending on the selected translation initiation rate.
The results will contain the pre-sequence, the synthetic ribosome binding site sequence, and the protein coding sequence. The translation initiation rate of the synthetic ribosome binding site / start codon and of any other start codons in the mRNA sequence will be displayed. Due to RNA-RNA interactions between the ribosome binding site and protein coding sequence, if you change the protein coding sequence, you should also design a new synthetic ribosome binding site.
Forward Engineering Mode with Constraints
Sometimes, it is necessary to insert a specific sequence within or nearby the ribosome binding site (e.g. a restriction site or a primer binding site). You can design a synthetic ribosome binding site and include such sequences by adding a sequence constraint.
Step 1: Add a sequence constraint in the RBS Constraints box. Enter any combination of ATCGU nucleotides to keep a nucleotide constant. Enter an N to allow a nucleotide to change. For example, the sequence constraint TCTAGANNNNNNNNNNNNNNN will generate a 21 nucleotide synthetic RBS sequence that begins with the XbaI restriction site.
Step 2: Similar to the forward engineering mode, specify the protein coding sequence, pre-sequence, and target translation initiation rate. Verify that the length of the pre-sequence and RBS sequence constraint adds up to at least 50 nucleotides.
Step 3: Click on the "Submit Job" button. The result will be stored in your "My Results" list and will be completed within 1 to 45 minutes, depending on the selected translation initiation rate and RBS Constraints. A solution may not exist if a a short or highly constrained RBS sequence is entered and/or a very high translation initiation rate is selected.
Reverse Engineering
Step 1: Copy and paste the sequence of an mRNA transcript into the mRNA Sequence box. The sequence should contain at least 50 nucleotides before and after the start codon of your protein(s) of interest. If a start codon is located near the beginning (5' end) of the mRNA transcript, only include the nucleotides in the mRNA transcript. Any combination of ATCGU nucleotides is acceptable. Thymines (T) are automatically converted to uracils (U).
Step 2: Click on the "Submit Job" button. The results will be stored in your "My Results" list and will be completed in less than 5 minutes.
The results will contain a list of start codons in the mRNA transcript and their predicted translation initiation rates on a relative scale from
Thank you, Saleh. I have a project that need to do a gene synthesis. i want to optimize the sequence before start the synthesis. But do you have experimental experience on using Forward or reverse engineering in RBS calculator?
I've used it several times before (different versions). It's great for finding and eliminating unwanted RBS from de novo gene synthesis projects. Hitting close to a particular absolute expression level is difficult, but in my experience usually if the Salis lab calculator predicts higher or lower relative expression, this will almost always be the case. Of course, my growth conditions are different than those used to refine this model, so I don't expect the same results at all, and the relevant papers don't claim perfect predictions (IIRC, measurements usually fall within 50-200% of the predicted value, but the model's been improved recently).
Sometimes for making very efficient RBS, I will directly look at RNA folding predictions for the 5' 100 bp or so and eliminating predicted secondary structure by mutating the 5' UTR and some codons.
Does the predicted RBS exit in real bacteria host? Or it is just the theoretical sequence? In this case, how can it work in real bacterial cells? Thank you for helping me clarify.
I've only used it for expression in E. coli (K12/MG1655). I've not checked whether predicted alternative RBSs exist (generally other ATG codons in or out of frame with my intended CDS). I used the RBS calculator to (1) eliminate predicted background RBSs, (2) design RBS efficiency to try to match that of some other gene from the same promoter, (3) make existing RBSs more or less efficient and (4) design a maximally efficient RBS de novo for some CDS. It's been at least partially successful for all of in E. coli.
@John - Not that I know of, and the Salis model won't work for eukaryotes. Eukaryotic RBSs differ in terms of RBS-to-start spacing, and translation initiation in eukaryotes is poorly matched by the model used here (basically 5' secondary structure + binding affinity for 16S ribosome + RBS-start spacing). Particularly, prokaryotic translational efficiency is largely determined by 16S RNA / mRNA (Shine Delgarno) complimentarity. The eukaryotic equivalent (Kozak sequence) doesn't work the same way (see link).
Ribosome is a ribonucleoprotein with rRNA and the sequence in the mRNA which needs to be expressed in the bacterial system will have the RBS (or atificially introduced mRNA) will have an RBS which will be complementary to the ribosomal 16s rRNA. This facilitates stable binding of the ribosome with the mRNA resulting in stable translation.
the study in attachment , researchers developed an equilibrium statistical thermodynamic model to quantify the strengths of the molecular interactions between the 30S complex and an mRNA transcript and to predict the resulting translation initiation rate.
I've used it only in reverse engineering mode, trying to understand the cause behind significant differences in expression levels between almost identical constructs, or as a guidance in picking what expression vector to use with what gene. For this purpose it is useful enough that it has become a part of my design routine, although 'natural' RBSs -non de novo-engineered- frequently throw off the predictions due to cases not accounted for by the model (overlapping RBSs or initiation codons, slow-folding sequences, etc).