Liang, N.; Li, J.-Y.; Ding, Y.; Wang, Y.; Champer, J.; Ren, L.-C.; Gao, G.

The 3\' region of genes plays a crucial role in transcriptional and post-transcriptional regulation, yet the functional elements within it remain largely unknown. Here, we assessed the impact of over 1.6 million synthetic 3\'UTR sequences on transcript abundance and subcellular localization in A549 cells, revealing a substantial number of relatively independent elements with distinct regulatory associations. By training a context-aware deep learning model, we found that these regulatory elements exhibit modest context dependency but possess unique preferences individually. Through relatively weak interactions and appropriate context arrangements, multiple elements can synergistically generate significant regulatory effects. Finally, we demonstrated the possibility for applying our model to identify disease-related 3\'UTR variants.