Variables (metabolites) are extracted that discriminates between the sample classes (the method is hence called OPLS-DA).
The sample size was hence 294.
However, many of these studies were based on small convenience samples, and were, hence, statistically underpowered and potentially unrepresentative.
Out of the selected 13 miRNAs, five miRNAs (hsa-miR210, hsa-miR296-5p, hsa-miR106-3p, hsa-miR34a-5p, and hsa-miR130b) showed amplification on qRT-PCR in <20% of the samples and were hence excluded from further analysis.
A corresponding UV-selected sample is under construction: hence, we only have an IR-selected sample.
As the sample is small and hence the power of the regression is low (power calculations made after the regression analysis ranged between 0.06 and 0.80) the interpretation of the non significant results must be done with care.
The latter is used to complete the representative learning dictionary by generating a theoretical pyrosequencing signal for each mutation for which no biological sample is yet available; hence, experimental signal is still lacking in the dictionary.
As mentioned earlier, the study sample is not representative, and hence, the generalizability and external validity of the findings are questionable.
For negative mode samples, computation time was hence below 0.5 min (Windows 7, R version 3.1.3, 2.2 GHz Intel core i7-4702 MQ processor, single-core usage, 32 GB RAM, 64 bit).
Temporal accuracy was set by sampling rate and was hence of 0.02 s.
