Abstract Acute lymphoblastic leukemia (ALL) is characterized by the presence of specific rearrangements associated with distinct gene expression signatures defined by high density microarray. Rearrangements of MLL, E2A/PBX1, BCR/ABL and TEL/AML1 are examples of this phenomenon, which also provide insights into the mechanisms of malignant transformation in these cases. Nevertheless, many cases of ALL do not carry known molecular abnormalities and do not exhibit unique gene expression profiles. In these cases, the mechanisms of malignant transformation remain unknown. Our goal was to develop an integrated genetic approach to discover new mechanisms of transformation in ALL and to define patterns of expression associated with these genetic abnormalities. Gene expression profiles were determined in B-lineage ALL from 95 adult patients using Affymetrix U95Av2 GeneChips. Within this group, high density SNP analysis was performed on 16 paired samples of normal and leukemia cells from the same patient using Affymetrix Mapping 10K Arrays. SNP analysis of normal and ALL DNA based on chromosome copy number interference revealed the presence of several regions of amplification, the most frequent involving chromosome X (3 samples, plus 1 Klinefelter syndrome, minimal amplified region: Xp22.31 to Xp11.1) and chromosome 21 (3 samples plus 1 Down syndrome, minimal amplified regions 21q223.13 to 21q22.2). Further analysis showed a frequent occurrence of LOH on chromosome 3, 6, and 21. Remarkably, this analysis revealed the presence of LOH in a portion of chromosome 9p (9p13.3 to 9p24.3) in 25% of cases. FISH analysis confirmed the presence of a homozygous deletion at 9p21 in 2 of these samples, whereas the 2 remaining cases showed a normal copy number, suggesting that in these 2 patients a deletion followed by duplication occurred. Analysis of gene expression based on genes located in this deleted region identified two genes with reduced expression: tropomyosin 2, a putative tumor suppressor gene and a transcript whith an unknown function. Analysis of gene expression in these cases revealed differential expression of 124 genes: of these, only 25 were highly expressed in the “9p LOH cases”. The use of this gene set in a heat map of all the cases without molecular abnormalities revealed a cluster that included 10 cases. 5 have known 9p abnormalities involving p15 and/or p16, 2 have 9p LOH by SNP analysis and 2 did not have sufficient material for further analysis. Given these results, we evaluated to what extent there was an overlap between these two phenomena. This revealed 34 transcripts that were consistently selected, suggesting that these genes represent a unique signature reflecting leukemic transformation involving the 9p region. These results demonstrate that integrated analysis of gene expression and SNP-based comparison of leukemia and normal cells can identify previously unknown groups of tumors with distinct genetic profiles. Further studies can now be undertaken to better define the mechanisms of transformation in these leukemias.