Recent studies have demonstrated that 3D genome alterations play important 26roles in tumorigenesis 1-3 , including the development of hematological 27 malignancies 4-7 . However, how such alterations may provide key insights into 28 T-lineage acute lymphoblastic leukemia (T-ALL) patients is largely unknown. 29 Here, we report integrated analyses of 3D genome alterations and differentially 30 expressed genes (DEGs) in 18 newly diagnosed T-ALL patients and 4 healthy T 31 cell controls. We found that 3D genome organization at the compartment, 32 topologically associated domains (TAD) and loop levels as well as the gene 33 expression profiles could hierarchically classify different subtypes of T-ALL 34 according to the T cell differentiation trajectory. Alterations in the 3D genome 35 were associated with nearly 45% of the upregulated genes in T-ALL. We also 36 identified 34 previously unrecognized translocations in the noncoding regions 37 of the genome and 44 new loops formed between translocated chromosomes, 38 including translocation-mediated enhancer hijacking of the HOXA cluster. Our 39 analysis demonstrated that T-ALLs with HOXA cluster overexpression were 40 heterogeneous clinical entities, and ectopic expressions of the HOXA11-A13 41 genes, but not other genes in the HOXA cluster, were associated with immature 42 phenotypes and poor outcomes. Our findings highlight the potentially 43 important roles of 3D genome alterations in the etiology and prognosis of 44 T-ALL. 45 Keywords: 3D genome alterations in T-ALL, chromosomal translocation, enhancer 46 hijacking, ectopic HOXA11-A13 expression, poor prognosis of T-ALL. 47 48 alterations that affect normal T cell development 8 . Recent whole-exome and RNA 49 sequencing analyses of large T-ALL cohorts are focused on the coding region of the 50 genome and have identified novel driver mutations, dysregulated transcription factors 51and pathways in T-ALLs 9-12 . To determine whether alterations in the 3D genome 52 organization are associated with malignant transformation of T-ALL, we conducted 53 BL-Hi-C 13 analysis using purified primary leukemic blasts from 18 newly diagnosed 54 T-ALL patients, including 8 early T-cell precursor ALL (ETP ALL) and 10 non-ETP ALL, 55 two clinical subtypes of T-ALL, as well as normal T cells from 4 healthy volunteers 56( Supplementary Fig. 1a). The maximum resolutions of the chromatin contact maps for 57 ETP, non-ETP ALL and normal samples were approximately 3.5, 3.5 and 10 kb, 58 respectively ( Supplementary Table 1). 59Principal component analysis (PCA) at the levels of the compartment, TAD and 60 loop structures demonstrated that the T-ALL samples could be separated from the 61 control samples by PC1, while ETP and non-ETP ALL could be separated by PC2 at 62 all three architectural levels ( Fig. 1a, upper panels) and be further delineated by 63 hierarchical clustering analysis ( Fig. 1a, lower panels). Detailed comparisons of the 64 3D chromosomal organizations of the T-ALL patients and the healthy controls 65 revealed 1.59% of ...