1) Protein-RNA docking - the field is in infancy, docking more rigid RNAs that do not change much after binding to the protein may succeed, but flexible RNAs, even small, would most likely fail. If you google a bit, there is a considerable information about protein-RNA docking already. For example, you easily can find that protein-RNA docking has been included in CAPRI (molecular docking contest) and it failed when unbound RNA structure was used.
2) RNA modification - if you want to oxidize nucleotides, you can try a tool from my previous lab: http://genesilico.pl/moderna/ "ModeRNA is capable of handling 115 different nucleotide modifications"
3) Protein in silico mutagenesis - I would use Rosetta http://www.rosettacommons.org/software/ together with http://www.pyrosetta.org/
4) There is little hope that your experiment will give any conclusive results - as I understand you want to dock RNA to protein and check how little changes to protein or RNA affect the result. I think that uncertainties in the docking results themselves are too big to analyze the effect of fine changes like single point mutants etc. The uncertainties are even bigger as it's still not straightforward to predict the effects of mutations on the structure.
I was docking dsDNA a lot, both by comparative modeling and de novo docking, and even in comparative modeling (much more accurate than de novo modeling) there were significant changes in several prot-DNA interactions due to even slight changes in DNA angles. No hope that in silico I could assess quantitatively the effects of single point mutants.
In one of the projects (http://www.ncbi.nlm.nih.gov/pubmed/17407166), I was also trying to explain the known effects of small DNA modifications based on protein-DNA model, and only some explanations were correct as revealed later by crystal structure (http://www.ncbi.nlm.nih.gov/pubmed/17344322)
Thanks for your explanations. That's a bit disappointing, but I'll try to find the way to do it in silico (maybe using machine learning methods based on existing data?) nonetheless, since I have no access to X-ray or NMR in Gliwice.
Start with analyzing well known structures of your protein domains in complex with RNA. Perhaps you can find some rules, regular patterns of binding modes, etc. Perhaps the complexes are so well characterized that you don't need docking at all.
And define what answers you expect to get in the best case, and what anybody could do with that data. It's easy to do a project in bioinformatics, more difficult is to provide any useful answers.
one of the main problems here is that no method treats efficiently macromolecular flexility for the whole molecule (i.e. auto dock)