For the read coverage correlation analysis, we used a similar approach to that in [ 9] to perform pair-wise comparisons for all of our data sets [Additional file 2: Figure S1 and Additional file 5: Table S3].
Additional file 1: Statistics of RNA-seq, normalized total gene read counts, correlations between samples and expressed genes lists.
Additional file 9: Statistics of RNA-seq, normalized total gene read counts, correlations between samples, list of DE genes and Gene Ontology enrichment for comparison between each sample, mean expression profile for superclusters and expression profiles of gene groups for second experiment.
The regression analysis showed that a linear adjustment to the data explains 85% of the CTL variance (R = 0.92, F 2,9) = 25.8, p<0.0002) with a significant contribution of the difference between paths for pseudoword reading (partial correlation t 9) = −5.36, p = 0.0004, R2 = −0.87) but not word reading (partial correlation t 9) = 1, p = 0.3, R2 = 0.30).
When including only orthologous pairs where both members had a minimum number of reads, the correlation coefficient remained quite stable.
Often used to analyze thermometer reading accuracy, correlation is an estimate of how much two variables change in relation to one another.
We found that small RNA level was positively correlated with mRNA level, but only when mRNA expression was above a clearly defined threshold of around 1,000 normalized reads (Spearman Correlation Coefficient > 0.75, Fig. 3a, Additional file 6).
The workflow for correlation analysis (List_CorrData) allows the user to calculate either Pearson or Spearman rank correlations or read in previously calculated correlation data.
The results showed different correlations between intrinsic value and the number of slides read, with positive correlations among high performers and negative correlations among low performers.
To test for gene expression correlations, read mapping counts obtained with Novoalign for all 63,683 reference transcripts were filtered removing in a first step transcripts expressed in less than 10% of all samples and subsequently with less than 1 million reads in fewer than two samples.
