Beyond Static Screens: A High-Throughput Pooled Imaging CRISPR Platform for Dynamic Phenotype Discovery

ABSTRACT Genetic screens are powerful tools for linking gene perturbations to cellular phenotypes, yet most pooled imaging approaches rely on static endpoint measurements. Here, we extend optical pooled screening to dynamic signaling phenotypes by combining time-lapse Förster resonance energy transfer–fluorescence lifetime imaging microscopy (FRET-FLIM) with optical enrichment and downstream genotyping. To enable high-throughput live-cell screening of dynamic phenotypes, we optimized and benchmarked all major steps in the pipeline. We developed FAST-HIPPOS, a Fiji-based automated analysis pipeline that rapidly segments hundreds of thousands of cells, extracts single-cell lifetime time traces, and identifies phenotypic hits in minutes for targeted photoactivation. We further established a robust phototagging strategy using the photoactivatable dye PA-JaneliaFluor646, providing bright, cloning-free labeling across diverse cell types. Benchmarking further demonstrated high sorting fidelity with low false-positive rates and efficient recovery of guide RNAs from as few as 200 sorted cells with minimal PCR bias. Applying this platform to cAMP signaling in HeLa cells using a custom CRISPR library recovered the expected primary regulators of β-adrenergic signaling—ADRB2, GNAS, and ADCY6—with no gene-level false positives. Together, these optimizations establish a robust general framework for screening dynamic phenotypes that is easily adaptable to other live-cell readouts beyond FRET-FLIM and cAMP. Competing Interest Statement The authors have declared no competing interest. Footnotes Dear reader, We would like to inform you that we have posted a revised version of the manuscript 'Beyond Static Screens: optical pooled screening of signaling dynamics using time-lapse FLIM' meant to replace the original upload with the title 'Beyond Static Screens: A High-Throughput Pooled Imaging CRISPR Platform for Dynamic Phenotype Discovery' on bioRxiv. These changes were made in part in response to feedback from an anonymous reviewer, who pointed out that version 1 failed to convey clear conceptual framing and presentation of the key advances. The original preprint posted on 11th July and revised on 31st July, 2025 contained the same experimental work but presented in a chronological narrative and, in retrospect, overly elaborate manner. The current revised version retains all experiments and conclusions, but has been substantially improved in terms of clarity, structure, and accessibility. In particular, we have streamlined the narrative, sharpened the central conceptual advance (screening of dynamic signaling phenotypes), and reorganized the Results and Discussion to more clearly separate methodological innovations from biological validation. Several technical sections have been condensed or moved to and from the supplementary materials, figures have been simplified, and the overall presentation has been made more concise and readable. To ensure a cohesive narrative, we have integrated new performance data for FAST-HIPPOS using the PA-JF646 dye. This experiment replaces a previous iteration involving PA-mCherry, providing a more consistent storyline. Additionally, we have removed data regarding "secondary hits" to maintain a focused narrative on the primary regulators that support our core conclusions. All other revisions are entirely focused on improving communication and interpretability, and no conclusions were altered. We believe that the revised version more effectively conveys the significance of the work and will be more accessible to a broad readership. Sincerely, Kees Jalink (on behalf of all authors)