TrendBio

#2 · Single-cell

New computational methods for single-cell, spatial, and imaging data analysis

6 preprints · 9 published · ↑ 2.4× vs 4-wk avg
Items per day, last 8 weekstoday: 15

The dominant thread is computational extraction of structure from single-cell and spatially resolved data, with recurring emphasis on integration, annotation, and benchmarking. Palette addresses mosaic integration of multimodal single-cell data with imbalanced modality composition, while scPathOT applies pathway-informed universal domain adaptation to harmonize scRNA-seq datasets across platforms and conditions for cell-type annotation. Lense uses an LLM to select preprocessing pipelines by comparing low-dimensional embeddings. Spatial methods include SpaNECT for niche characterization across spatial multimodal and multi-omics data and minkiPy for Minkowski-based profiling of spatial gene organisation. scShapeBench introduces a benchmark for geometric shape detection in scRNA-seq data with a Reeb-graph baseline. Other items extend analysis to live-cell fluorescence traces (SPIFEE), morphology-based inference of neuroblastoma cell states, Xenopus morphodynamic fate prediction, and volume EM vesicle segmentation. Newer contributions emphasize automated pipeline selection and explicit benchmarking.

Published papers

Preprints

Topic history

DateRank
#2
#1
#3
#3
#1
#4
#1
#2