Periodic Reporting for period 4 - DECODE (Decoding Context-Dependent Genetic Networks in vivo)
Okres sprawozdawczy: 2024-01-01 do 2025-06-30
The evolutionary success of multicellular organisms is based on the division of labor between cells. While some of the molecular determinants for cell fate specification have been identified, a fundamental understanding of which genetic activities are required in each cell of a developing tissue is still outstanding. The DECODE project develops and applies leading-edge system genetics methods to Arabidopsis and Drosophila, two major model systems from the plant and animal kingdoms to decode context-dependent genetic networks in vivo. To achieve this, DECODE brings together experimental and theoretical groups with complementary expertise in model organism genetics and cellular phenotyping, single-cell genomics, statistics and computational biology. Building on our combined expertise, we create functional genetic maps using conditional CRISPR-based single- and higher order knockout perturbations in vivo combined with single-cell expression profiling and imaging. Coupled with powerful computational analysis, this project will not only define, predict and rigorously test the unique genetic repertoire of each cell, but also unravel how genetic networks adapt their topology and function across cell types and external stimuli. Analyzing two tissues with divergent organization and regulatory repertoire will enable us to uncover general principles in the genetic circuits controlling context-dependent cell behavior. Consequently, we expect that the DECODE project in model organisms will lay the conceptual and methodological foundation for perturbation-based functional atlases in other tissues or species.
The DECODE team developed and applied experimental and computational tools to systematically analyze genetic network systems in vivo on a large scale never seen before. For the experimental approaches, we are using Drosophila and Arabidopsis as two major model systems from the animal and plant kingdoms.
We generated and improved transgenic in vivo CRISPR libraries for Drosophila to knockout genes in the whole fly or in selected tissues by gene editing. Subset screens targeting signaling pathway components in intestinal stem cells revealed pathway cross-talks and identified candidates that we are currently following up. We developed a new method to directly detect CRISPR mutations and transcriptional responses in the same cell, enabling us to track mutations across different lineages and understand gene programs involved in cell type formation. Combinatorial genetic perturbations coupled to single cell RNA-seq contributed to our understanding how genetic interactions impact different cell types along the gut and to study genotype-phenotype interactions in vivo. For Arabidopsis we developed and established an inducible CRISPR-Cas mediated gene knockout system and generated a library of knock-out mutant transgenic lines. We optimized perturbation and single-cell sequencing strategies, which was more difficult than foreseen originally. We were able to generate sets of experimental data focusing on perturbation of specific signaling pathways. The generation and experimental analysis of perturbation lines was continuously expanded and combinatorial index sequencing on Arabidopsis roots was established.
To exploit generated data, we developed a scalable and robust data management infrastructure including workflows for data processing and downstream biological analyses. We implemented an automated pipeline to visualize preliminary analyses. Our data management strategy demonstrated its utility in our daily work, facilitating data sharing and fostering collaborative work across the groups. In parallel, we developed, benchmarked and improved foundational single-cell data analysis tools. We established a scalable workflow to process and analyze sci-RNAseq data. As a core objective is to quantify phenotypic changes across molecular layers and model systems, we further developed computational methods beyond state-of-the-art to analyze high-throughput, multidimensional phenotype perturbation data and to integrate multi-omics single-cell profiles, with temporal and spatial context. We developed a novel computational framework enabling the analysis of multi-condition datasets without or prior to categorizing cells into discrete cell types. We implemented it into a software package, LEMUR, which now enables us to systematically disentangle shared and unique perturbation effects while accounting for complex, hierarchical or gradual cell type composition in in vivo samples. Currently, we made advances in the development of computational methods to better link CRISPR perturbations with single-cell transcriptomic readouts, including the development of technologies for direct mutation sequencing and strategies to assay editing outcomes alongside gene expression.
Overall, the project progressed very well, the results achieved so far have been published, and we are continuing to work on the large-scale screening approaches and the systematic analysis of context-specific genetic networks.
We generated and improved transgenic in vivo CRISPR libraries for Drosophila to knockout genes in the whole fly or in selected tissues by gene editing. Subset screens targeting signaling pathway components in intestinal stem cells revealed pathway cross-talks and identified candidates that we are currently following up. We developed a new method to directly detect CRISPR mutations and transcriptional responses in the same cell, enabling us to track mutations across different lineages and understand gene programs involved in cell type formation. Combinatorial genetic perturbations coupled to single cell RNA-seq contributed to our understanding how genetic interactions impact different cell types along the gut and to study genotype-phenotype interactions in vivo. For Arabidopsis we developed and established an inducible CRISPR-Cas mediated gene knockout system and generated a library of knock-out mutant transgenic lines. We optimized perturbation and single-cell sequencing strategies, which was more difficult than foreseen originally. We were able to generate sets of experimental data focusing on perturbation of specific signaling pathways. The generation and experimental analysis of perturbation lines was continuously expanded and combinatorial index sequencing on Arabidopsis roots was established.
To exploit generated data, we developed a scalable and robust data management infrastructure including workflows for data processing and downstream biological analyses. We implemented an automated pipeline to visualize preliminary analyses. Our data management strategy demonstrated its utility in our daily work, facilitating data sharing and fostering collaborative work across the groups. In parallel, we developed, benchmarked and improved foundational single-cell data analysis tools. We established a scalable workflow to process and analyze sci-RNAseq data. As a core objective is to quantify phenotypic changes across molecular layers and model systems, we further developed computational methods beyond state-of-the-art to analyze high-throughput, multidimensional phenotype perturbation data and to integrate multi-omics single-cell profiles, with temporal and spatial context. We developed a novel computational framework enabling the analysis of multi-condition datasets without or prior to categorizing cells into discrete cell types. We implemented it into a software package, LEMUR, which now enables us to systematically disentangle shared and unique perturbation effects while accounting for complex, hierarchical or gradual cell type composition in in vivo samples. Currently, we made advances in the development of computational methods to better link CRISPR perturbations with single-cell transcriptomic readouts, including the development of technologies for direct mutation sequencing and strategies to assay editing outcomes alongside gene expression.
Overall, the project progressed very well, the results achieved so far have been published, and we are continuing to work on the large-scale screening approaches and the systematic analysis of context-specific genetic networks.
To reach the project goals, the DECODE team developed resources, methodological procedures and novel strategies for both experimental and computational approaches. Here we would like to highlight the following achievements published in internationally recognized journals and made publicly available. The Drosophila in vivo CRISPR-Cas libraries (Port et al, PNAS 2020; Port et al, eLife 2020) opened up new possibilities for systematic screening approaches and were made available through the Vienna Drosophila Resource Center (VDRC). We further optimized methods to perform mutagenesis with spatial and temporal control in Drosophila. We also developed a new system based on the induction of premature stop codons with cytidine base editors (Doll et al, Science Advances 2023). Furthermore, we are focusing on increasing our perturbation throughput by combining methods. We developed and applied a novel single-cell methodology to generate one of the first single-cell RNAseq maps (Liu et al., Mol Plant 2021) as well as single cell chromatin accessibility map in a crop. For first-stage data analysis of single cell RNAseq data, we developed a new method that scales to the size of realistic datasets (Ahlmann-Eltze and Huber, Bioinformatics, 2020) and made it available as R/Bioconductor package (glmGamPoi). With LEMUR, we developed a novel computational framework that enables the analysis of multi-perturbational datasets without or prior to categorizing cells into discrete cell type (Ahlmann-Eltze and Huber, Nature Genetics, 2025). As an open-source package, it is already being used by other projects in diverse areas of science. We further developed novel computational strategies for assaying genetic effects at the single-cell level (Cuomo et al., Mol Sysbiol, 2022), as well advancing the computational foundations for integrating high-dimensional spatio-temporal data from perturbation screens (Velten et al, Nature Methods, 2022). With these achievements, along with the large experimental data-sets recorded, we are in a good position to successfully pursue the DECODE project aim to lay the conceptual and methodological foundation for perturbation-based functional atlases, to create the largest single-cell perturbation maps of the model organisms used and to provide fundamental insights into the genetic architecture of complex tissues.