I am attempting to use RNAcalibrate in order to predict extreme value distribution for some miRNA-target interactions which I intend to then pass to RNAhybrid. As a test I started with the c.elegans miRNA cel-let-7-3p and used dinculotide frequency data I obtained using a R script that analyzed the annotated c.elegans 3'UTR available from ensembl.
However, when I passed the parameters I obtained from RNAcalibrate to RNAhybrid (again querying cel-let-7-3p against all c.elegans 3'UTR), RNAhybrid generated an output in which all predicted targets have a p-value of 1.000. When I tried to run RNAhybrid with the same query and target sequence using RNAhybrid built-in c.elegans 3'UTR set, I did receive different p-values for different targets, indicating that the first output I received was not an artifact of the particular miRNA or target sequences.
When I tried running RNAhybrid against other sets of query and target sequences from other species with evd parameters I obtained using RNAcalibrate I got similar results (that is, all targets had the same p-value, although not necessarily 1.000).
As I indicated, I'm aware of the fact that RNAhybrid has a built-in evd for c.elegans, but I'm working in a project in which I need to generate miRNA target prediction for several species, some of which do not have built in target datasets in RNAhybrid. I started with c.elegans as a test in order to be able to compare the results I get using the evd calculated by RNAcalibrate with the result generated by the built-in target set.
Do you have any idea what the source of the error might be and how I can refine my usage of RNAcalibrate to get more meaningful results? Any help will be highly appreciated.
The command lines I used are provided below if you want to examine them.
Thanks in advance
RNAcalibrate -d celegans_3UTRS.freq -q cel-let-7-5p.fa > cel_let7_evd.txt
RNAhybrid -e -20 -d 1.661933 0.108887 -t celegans_3UTRS.fa -q cel-let-7-5p.fa > RNAhybrid_cel_let_7p_calib.txt