Abstract: Three sets of 20 trinucleotides are preferentially associated with
the reading frames and their 2 shifted frames of both eukaryotic and
prokaryotic genes. These 3 sets are circular codes. They allow retrieval of any
frame in genes (containing these circular code words), locally anywhere in the
3 frames and in particular without start codons in the reading frame, and
automatically with the reading of a few nucleotides. The circular code in the
reading frame, noted X, which can deduce the 2 other
circular codes in the shifted frames by permutation, is the information used
for analysing frameshift genes, i.e. genes with a change of reading frame
during translation. This work studies the circular code signal around their
frameshift sites. Two scoring methods are developed, a function
P based on this code X and a function
Q based both on this code X and the 4
trinucleotides with identical nucleotides. They detect a significant
correlation between the code X and the −1
frameshift signals in both eukaryotic and prokaryotic genes, and the
+1 frameshift signals in eukaryotic genes.
