Systemic lupus erythematosous (SLE) is an autoimmune disease with an important clinical and biological heterogeneity. B lymphocytes appear central to the development of SLE which is characterized by the production of a large variety of autoantibodies and hypergammaglobulinemia. In mice, immature B cells from spontaneous lupus prone animals are able to produce autoantibodies when transferred into immunodeficient mice, strongly suggesting the existence of intrinsic B cell defects during lupus. In order to approach these defects in humans, we compared the peripheral B cell transcriptomas of quiescent lupus patients to normal B cell transcriptomas. When the statistical analysis is performed on the entire group of patients, the differences between patients and controls appear quite weak with only 14 mRNA genes having a false discovery rate ranging between 11 and 17%, with 6 underexpressed genes (PMEPA1, TLR10, TRAF3IP2, LDOC1L, CD1C and EGR1). However, unforced hierarchical clustering of the microarrays reveals a subgroup of lupus patients distinct from both the controls and the other lupus patients. This subgroup has no detectable clinical or immunological phenotypic peculiarity compared to the other patients, but is characterized by 1/an IL-4 signature and 2/the abnormal expression of a large set of genes with an extremely low false discovery rate, mainly pointing to the biological function of the endoplasmic reticulum, and more precisely to genes implicated in the Unfolded Protein Response, suggesting that B cells entered an incomplete BLIMP1 dependent plasmacytic differentiation which was undetectable by immunophenotyping. Thus, this microarray analysis of B cells during quiescent lupus suggests that, despite a similar lupus phenotype, different biological roads can lead to human lupus.