I have identified some piRNAS from miRNA-seq, but I want to know if exists a database like mirwalk that contains a list of predicted and validated posible gene (or proteins) target interactions.
Somebody knows if it is possible??
It is different than miRNA target identification. piRNAs have a perfect match to the target sequence so identifying the genomic features to which your reads map to will give you the targets. You can do that using bedtools intersect for instance. Now there a few things to bear in mind:
piRNAs usually target transposons which have multiple copies in the genome, so think carefully about how many mapping positions per read the aligner outputs, and how those are accounted for. If you are only interested in protein-coding genes, using uniquely mapped reads is a good (safe) option.
Unlike miRNAs, piRNAs are variable length which is also slightly different according to your species. I would remove reads outside the expected range if piRNA lengths.
If you are interested in transposon targeting, consider using RepEnrich or similar tools which have been developed for this effect. piPipes would be my advise since you appear to be a novice in piRNA analysis. If your species of interest is mouse or human it is also easier to set-up. The paper and manual is also worth looking at since it details the analysis.