Hello everyone, I'm using miRDeep2 to predict novel miRNAs in a Differential Expresion experiment from some muscle samples.
The thing is that when running miRDeep2, everything goes fine, but the problem comes when analyzing the results.
Does anyone know how to interpret the miRDeep2 scores?. The program generates a table with a header which tells you that the greater the miRDeep2 score is, the greater the probability is that this novel miRNA es a false positive, so by this, I understand that what is valuable to predict a good novel miRNA is that this miRNA has a low mirDeep2 score, near 0 or negative.
BUT, when going into the list of novel predicted miRNAs, which are designated by their coordinates, I find that most, if not all, of Differentially expressed predicted novel miRNAs, appear to have very high miRDeep2 scores, so according to what I mentioned above, those should have a very high probability of being false positives. BUT, as a variable within this list of novel predicted miRNAs, there's a field telling you the "estimated probability that the miRNA candidate is a true positive" which tells you that the higher the miRDeep2 is, the higher this probability of being a true positive is.
I find it conflictive, as firstly miRDeep2 score seems to be desirable if having a low value, but then it turns to the inverse interpretation, as it should have a high value.
Did someone analyze results from miRDeep2? I'm trying to select some desirable novel miRNAs to validate them via rt-qPCR but this is blowing my mind...
The thing is that when running miRDeep2, everything goes fine, but the problem comes when analyzing the results.
Does anyone know how to interpret the miRDeep2 scores?. The program generates a table with a header which tells you that the greater the miRDeep2 score is, the greater the probability is that this novel miRNA es a false positive, so by this, I understand that what is valuable to predict a good novel miRNA is that this miRNA has a low mirDeep2 score, near 0 or negative.
BUT, when going into the list of novel predicted miRNAs, which are designated by their coordinates, I find that most, if not all, of Differentially expressed predicted novel miRNAs, appear to have very high miRDeep2 scores, so according to what I mentioned above, those should have a very high probability of being false positives. BUT, as a variable within this list of novel predicted miRNAs, there's a field telling you the "estimated probability that the miRNA candidate is a true positive" which tells you that the higher the miRDeep2 is, the higher this probability of being a true positive is.
I find it conflictive, as firstly miRDeep2 score seems to be desirable if having a low value, but then it turns to the inverse interpretation, as it should have a high value.
Did someone analyze results from miRDeep2? I'm trying to select some desirable novel miRNAs to validate them via rt-qPCR but this is blowing my mind...