Neoantigen mRNA vaccine designTCRshows
String several neoantigen peptides into a single open reading frame, pick a linker, add a signal peptide and MITD — and then do the step this kind of design most often skips: enumerate every 8–11mer created across the junctions, and reorder the peptides so that as few of those as possible carry an HLA anchor motif.
1Neoantigen peptides
One per line, optionally with a >name FASTA header. Personalised vaccines usually use 25–27 aa long peptides with the mutation centred; 8–11 aa short epitopes work too.
2Linker and terminal elements
The linker is the first gate on junctional epitopes: in a flexible GS linker, G and S are poor anchor residues for most HLA supertypes, which by itself makes junction peptides unlikely to bind well.
✓Results
What each linker does (same peptides, each with its own best optimised order)
Junction-spanning neopeptides (only those carrying an anchor motif; none of these sequences occurs in any input peptide)
Full-length open reading frame
Why junctions deserve their own step. In a string-of-beads construct, every pair of adjacent peptides creates sequence that is present in none of the inputs. Two kinds of risk follow. One is that a junction peptide becomes an immunodominant decoy, pulling the response away from the real neoantigens. The other is that it happens to match a self protein, which is an autoimmunity risk. The first is suppressed by the choice of linker and the peptide order; the second can only be settled by a proteome comparison. All results are for research use only.
Sources: the MITD and the MHC-I signal peptide are taken from the feature annotations of UniProtKB P04439 (HLA-A α chain) — signal peptide 1–24, transmembrane 309–332, cytoplasmic 333–365. The anchor-position motifs are the accepted preferences of the common HLA supertypes and are meant for qualitative screening only.