CRISPR screens of noncanonical coding sequences and integrative multiomics uncover novel retrovirus-derived proteins

Ribosome profiling has revealed thousands of noncanonical translation events across mammalian genomes, yet functional characterization has overwhelmingly focused on proliferative fitness in cancer cell lines. Here, we present a comprehensive survey of noncanonical translation in the mouse immune system and its functional consequences in macrophages. By performing a meta-analysis of Ribo-seq across 20 public mouse leukocyte and 5 human leukocyte datasets we define a compendium of 22,572 mouse and 18,701 human noncanonical coding sequences (nCDSs), including upstream ORFs (uORFs), downstream ORFs, and ORFs on noncoding RNAs and pseudogenes (ncORFs). Cross species conservation and proteogenomic integration with reanalyzed mass spectrometry data prioritizes a high-confidence subset. To move beyond cataloging, we carried out two orthogonal CRISPR screens in immortalized bone marrow-derived macrophages: a fitness screen identifying nCDSs required for macrophage viability, and a TLR1/TLR2-NF{kappa}B reporter screen uncovering nCDSs that modulate innate immune signaling. These screens suggested dozens of nCDSs for functional follow up including a family of endogenous retroviral envelope-derived proteins translated in adult mouse myeloid cells. Updated single-cell RNA-seq annotations and an interactive UCSC Genome Browser session integrating Ribo-seq, proteomics, and CRISPR screen data are provided as community resources. Together, these findings expand the functional landscape of noncanonical translation in immunity and nominate endogenous retroviral proteins as previously unrecognized regulators of murine macrophage biology.