We combine four independently generated datasets from 97 individuals to study CNVs for the first time at high-resolution in a Gulf Arab population.
This used public datasets from three studies that used the CEU population [16] [18], with each study using a different sub-cohort of individuals and different platform to measure expression levels.
It is common to combine data from individual expression studies into one large dataset prior to analysis.
To do that, they considered 29,000 datasets from 75 studies.
We used three different ESR datasets from previous studies [29], [30], [40], and a consensus dataset (consisting of ESR motifs contained in all three datasets, C-dataset).
In our previous study, we found that the ComBat algorithm could remove such noise (batch signal or each individual study) from the dataset with powerful efficiency when adjusted with additional and multiple effects of the batch information [ 15, 16], which provide the prerequisite to combine the methylation array dataset to increase the sample size of the statistical analysis.
In this study, three multicenter genome-wide transcriptomic datasets from 79 individuals were used to infer rule-based classifiers to discriminate RA, OA, and healthy controls.
Individual datasets from six prospective MIBG cohort studies performed in Japan between 1990 and 2009 were combined to make a pooled database of 1,322 CHF patients as previously reported [ 5, 13– 13].
In this study, a global view comparison of the complete dataset from each individual biological replicate showed that all datasets generated consistent and reproducible results, underlining the high quality iCLIP data achieved by high stringency purification and library preparation.
