No abstract

Spatial transcriptomics promises to greatly improve our ability to understand tissue organization and cell-cell interactions. While most current platforms for spatial transcriptomics only provide multi-cellular resolution (10-15 cells per spot), recent technologies provide a much denser spot placement leading to sub-cellular resolution. A key challenge for these newer methods is cell segmentation and the assignment of spots to cells. Traditional, image based, segmentation methods face several drawbacks and do not take full advantage of the information profiled by spatial transcriptomics. Here, we present SCS, which integrates imaging data with sequencing data to improve cell segmentation accuracy. SCS combines information from neighboring spots and employs a transformer model to adaptively learn the relevance of different spots and the relative position of each spot to the center of its cell. We tested SCS on two new sub-cellular spatial transcriptomics technologies and compared its performance to traditional image based segmentation methods. As we show, SCS achieves better accuracy, identifies more cells and leads to more realistic cell size estimation. Analysis of RNAs enriched in different sub-cellular regions based on SCS spot assignments provides information on RNA localization and further supports the segmentation results.

No abstract

Annotating the functions of gene products is a mainstay in biology. A variety of databases have been established to record functional knowledge at the gene level. However, functional annotations at the isoform resolution are in great demand in many biological applications. Although critical information in biological processes such as protein–protein interactions (PPIs) is often used to study gene functions, it does not directly help differentiate the functions of isoforms, as the ‘proteins’ in the existing PPIs generally refer to ‘genes’. On the other hand, the prediction of isoform functions and prediction of isoform–isoform interactions, though inherently intertwined, have so far been treated as independent computational problems in the literature. Here, we present FINER, a unified framework to jointly predict isoform functions and refine PPIs from the gene level to the isoform level, enabling both tasks to benefit from each other. Extensive computational experiments on human tissue-specific data demonstrate that FINER is able to gain at least 5.16% in AUC and 15.1% in AUPRC for functional prediction across multiple tissues by refining noisy PPIs, resulting in significant improvement over the state-of-the-art methods. Some in-depth analyses reveal consistency between FINER’s predictions and the tissue specificity as well as subcellular localization of isoforms.