Integration of retroviral cDNA into the host cell genome is a process central to the replication cycle of retroviruses and is mediated by the virally-encoded integrase protein. While any DNA sequence can be a target for integration, retroviruses do not integrate randomly into the host cell chromosomes. Recent studies have found that different retroviruses have distinct target site selection preferences for various genomic features. We have sequenced integration sites from human immunodeficiency virus (HIV), murine leukemia virus (MLV) and three HIV-MLV chimeras and determined that both integrase and the viral Gag proteins act together to determine virus-specific integration site selection preferences. Once integrated, the provirus is transcribed by the host cell machinery into messenger RNA and the viral RNA genome. A number of factors are thought to contribute to the level of proviral gene expression. For HIV, these include the activation state of the host cell, CpG methylation, nucleosome organization, and mutations in the viral transactivator, tat, or transcription factor binding sites in the viral promoter and enhancer. Factors that negatively influence HIV gene expression are of interest because of the phenomenon of viral latency, where HIV persists in the genome of host cells undetected due to a lack of expression. We set out to determine the extent to which integration site influences expression of the HIV provirus. In this study, we infected Jurkat T cells with an HIV-based vector transducing GFP and separated cells into GFP-expressing and non-expressing populations. We then sequenced integration sites from these two populations. Low proviral expression correlated with integration into (1) gene deserts, (2) centromeric heterochromatin, and (3) very highly expressed host cell genes. These data suggest that particular genomic features influence the expression of HIV proviruses and provide models for postintegration latency in cells from HIV-infected patients