Skip to main content
Fig. 2 | BMC Biology

Fig. 2

From: An integrated in silico-in vitro approach for identifying therapeutic targets against osteoarthritis

Fig. 2

Microarray integration for GRN inference. A Assembling and correcting the microarray sub-datasets. PCA plots and gene expression distributions of the assembled dataset colored by arrays (GSE-number), before and after quantile normalization and batch effect correction with ComBat. B Unsupervised clustering with the Euclidian complete method highlights that the samples do not cluster according to the technological platform but rather according to the OA status of the samples. When splitting the hierarchical tree into three branches, a mostly OA group and a mostly WT group stand out. The table summarizes the percentage of true OA (resp. WT) samples correctly grouped in the “OA-group” (resp. “WT-group”) highlighting a clustering accuracy in line with what is obtained for the individual sets before correction. C the algorithms compute possible regulatory interactions and output a list of possible transcriptional interactions from a transcription factor to another gene. “g1,” “g2,” and “g3” denote gene1, 2, and 3. D List of interactions inferred with the merged OA dataset and integrated into the mechanistic model (11 predictions in total). Inference was run with three algorithms, solely interactions that were present in the results of the three algorithms were reported and integrated. An interaction was considered present for one algorithm if its score was higher than a threshold defined as the difference between the mean and standard deviation of all scores (Additional file 6, Data S2). Corr.score is the Spearman correlation coefficient, computed solely to define the interaction sign (activation if positive, inhibition if negative). The validity of inferred interactions was supported by looking for binding sites of the source gene in the enhancer region of the target with using the GeneHancer database embedded in GeneCards [26]. GeneHancer IDs are reported, when found. ‘etc.’ in indicated when more than one was found. The queried gene IDs and the full list of GeneHancer IDs are reported in Additional file 7, Table S3

Back to article page