Lung injury after pediatric allogeneic hematopoietic cell transplantation (HCT) is a common and disastrous complication that threatens long-term survival. In order to develop strategies to prevent lung injury, novel tools are first needed to comprehensively assess lung health in HCT candidates. Therefore, this study analyzed biospecimens from 181 pediatric HCT candidates who underwent routine pre-HCT bronchoalveolar lavage (BAL) at the University Medical Center in Utrecht, the Netherlands between 2005-2016. BAL fluid underwent metatranscriptomic sequencing of microbial and human RNA and unsupervised clustering and generalized linear models were used to associate microbiome-gene expression data with the development of post-HCT lung injury. Microbe-gene correlations were validated using a geographically distinct cohort of 18 pediatric HCT candidates. The cumulative incidence of post-HCT lung injury varied significantly according to four pre-HCT pulmonary metatranscriptome clusters, with the highest incidence observed in children with pre-HCT viral enrichment and innate immune activation and in children with profound microbial depletion and concomitant NK/T-cell activation (p<0.001). In contrast, children with pre-HCT pulmonary metatranscriptomes containing diverse oropharyngeal taxa and lacking inflammation rarely developed post-HCT lung injury. In addition, activation of epithelial-epidermal differentiation, mucous production, and cellular adhesion were associated with fatal post-HCT lung injury. In a separate validation cohort, associations between pulmonary respiratory viral load, oropharyngeal taxa, and pulmonary gene expression were recapitulated; the association with post-HCT lung injury needs to be validated in an independent cohort. This analysis suggests that assessment of the pre-HCT BAL fluid may identify high-risk pediatric HCT candidates who may benefit from pathobiology-targeted interventions.