Fig. 2

Multiple samples clustering. Hierarchical clustering constructed with all OR FPKM log values, that correspond to the number of times each transcript was sequenced, which itself depend on their concentration in the libraries (Additional file 3). Most of the samples from the same animal are in the same cluster, the main exception are the two Belgian Shepherd samples, one of them being clustered with the Labrador Retriever samples, the second being alone. This overall good clustering indicates the good reproducibility of the sampling and of the analysis. The distribution of the two Belgian Shepperd samples in two different clusters suggests a pathological problem [30]