InosinePredict uses an algorithm based on an experimentally determined in vitro editing of a long, perfectly base-paired dsRNA, as described in:

Predicting sites of ADAR editing in double-stranded RNA
Julie M. Eggington, Thomas Greene, Brenda L. Bass, 2011 Nat Commun 2:319.

Predictions take into account the identity of the four bases 5’ and four bases 3’ of an adenosine, and are based on experimentally determined editing site preferences for human ADAR1, human ADAR2, and proteins truncated to only contain the catalytic domain (hADAR1-D and hADAR2-D).


To apply the algorithm to your RNA sequence, upload a file in FASTA or Multi-FASTA format OR cut and paste your FASTA sequence in the box provided in step 1.  Observe the following:

Step 1

Cut and paste your FASTA formatted sequence(s) here:


Upload a file in FASTA format:    

Step 2

Select an ADAR neighbor preference table to predict editing.

Please make a selection.




InosinePredict will predict editing sites in the RNA sequence uploaded, as well as the complementary strand, assuming Watson-Crick base-pairing. At present, GU base-pairs are not considered, but can be accounted for by entering each strand separately.

InosinePredict requires at least four neighboring bases on each side of the relevant adenosine. Adenosines less than 4 nucleotides from 5’ or 3’ termini are not considered.