Biology subjects

Richter, T.

Publications and source records attributed to Richter, T..

11 recordsLinked to original sources

bioRxiv2026

PIGSTI: a modular, reproducible pipeline for detecting species identity, pathogens, and microbes from animal palaeogenomic data

Ancient genomics has enabled discovery of diverse pathogens across various time periods, host species, and material types. However, existing palaeogenomic pipelines predominantly focus on screening data from human hosts, or do not incorporate microbial screening methodologies. We present PIGSTI (Pathogen anImal Genome Sequence ToolkIt), a bioinformatic pipeline specifically designed for both the initial screening and subsequent detection of pathogens in shotgun sequencing data from ancient animal remains. PIGSTI's integrated Snakemake workflow performs both host detection, genome mapping and pathogen identification, generating outputs suitable for population genetics and phylogenetic analyses. Testing on 952 newly sequenced and publicly available animal palaeogenomic datasets, we identified ~15 ancient zoonotic and animal pathogens with high confidence, including the first documented case of Rickettsia felis and Leptospira borgpetersenii in an ancient animal. Our results demonstrate PIGSTI's utility for screening pathogen diversity in ancient animal hosts and reconstructing historical host-pathogen relationships.

genomics↗

bioRxiv2026

Leukemic Stem Cell Subtypes Drive Distinct Niche Remodeling in Acute Myeloid Leukemia

Leukemic stem cells (LSCs) sustain acute myeloid leukemia (AML) and are implicated in therapy resistance and relapse. Yet, it remains unknown how LSCs remodel the bone marrow niche. AML is known to alter stromal and vascular microenvironments, but these effects are difficult to separate from bulk leukemic burden and immune inflammation. Here, we use isogenic human AML xenografts with distinct LSC characteristics but comparable engraftment to define LSC-associated niche remodeling in vivo. Single-cell profiling revealed that LSC-high AML shifts the mesenchymal niche toward fibro-inflammatory states, expanding Fmod+ fibroblasts and Cd34+ perivascular fibroblast-like cells while suppressing osteolineage differentiation. The leukemic compartment remained heterogeneous, with a specific MEP-like LSC population expressing niche-remodeling ligands including TGFB1, IL1B, and ANGPT1. LSC-high AML activated a TGF{beta}-responsive, CREB3L1-controlled fibroblast trajectory, and perturbing TGF{beta} signaling or CREB3L1 activation reduced stromal support for AML cells. These findings identify a specific LSC subtype as a source of niche-remodeling cues that shape specialized leukemia-supportive niches.

molecular biology↗

bioRxiv2026

Perturbation-guided mapping of colorectal cancer cell states to causal mechanisms

Colorectal cancer (CRC) cell atlases have refined descriptive maps of tumour ecosystems, yet cross-sample integration often obscures disease-relevant patient-specific variation and remains largely correlative, limiting insight into the mechanisms and state transitions that drive progression and treatment response. Here, we develop a continual learning framework to construct a comparative single-cell CRC atlas spanning over 300 patients and 1.5 million cells, preserving inter-patient variation while aligning healthy and malignant contexts. We resolve distinct non-canonical malignant cell states, including an endoderm-like state enriched in microsatellite-stable, KRAS-mutant CRC with features of oncofetal plasticity. Cell states are recapitulated in patient-derived organoids, establishing a tractable model of reprogramming. By linking the observational atlas to a large-scale perturbation atlas using relative representations, we map perturbations that drive cells toward defined phenotypic extremes. We connect cell states to therapeutic responses, showing that MAPK inhibition induces a shift away from a proliferative phenotype and converges towards a plastic, endoderm-like state. Together, this framework moves beyond static atlases to enable mechanistic modeling of cell-state regulation and causal inference toward cell-state-directed therapies.

genomics↗

bioRxiv2026

Beyond alignment: synergistic integration is required for multimodal cell foundation models

The vision of a "virtual cell"--a computational model that simulates biological function across modalities and scales--has become a defining goal in computational biology. While powerful unimodal foundation models exist, the lack of large-scale paired data prohibits the joint training of multimodal approaches. This scarcity favors compositional foundation models (CFMs): architectures that fuse frozen unimodal experts via a learned interface. However, it remains unclear when this multimodal fusion adds task-relevant information beyond the strongest unimodal representation and when it merely aggregates redundant signal. Here, we introduce the Synergistic Information Score (SIS), a metric grounded in partial information decomposition (PID), that quantifies the information gain achievable only through cross-modal interactions. Extending theoretical results from self-supervised learning, we show that standard alignment-based fusion objectives on frozen encoders inherently collapse to detecting linear redundancies, limiting their ability to capture nonlinear synergistic states. This distinction is directly relevant for tasks aiming to link tissue morphology and gene expression. Benchmarking ten fusion methods on spatial transcriptomics datasets, we use SIS to demonstrate that tasks dominated by linear redundancies are sufficiently served by unimodal baselines, whereas complex niche definitions benefit from synergy-aware integration objectives that enable cross-modal interactions beyond linear alignment. Finally, we perform a scaling analysis which highlights that fine-tuning a dominant unimodal expert is the most sample-efficient path for standard tasks, suggesting that the benefits of multimodal frameworks only emerge when tasks depend on information distributed across modalities. Together, these results establish that building towards a virtual cell will require a fundamental shift from alignment objectives that emphasize shared structure to synergy-maximizing integration that preserves and exploits complementary cross-modal signal.

bioinformatics↗

bioRxiv2025

Language may be all omics needs: Harmonizing multimodal data for omics understanding with CellHermes

Decoding cellular systems requires integrating diverse omics data, yet most models are trained from scratch on a single modality, restricting generalization. Here we present CellHermes, a biological language model that leverages pretrained large language models (LLMs) to integrate multimodal forms of omics data, such as transcriptomic profile and PPI network, through natural language for better omics understanding. By reformulating these datasets into question-answer pairs, CellHermes emulates multiple self-supervised learning paradigms within a single, universal, language-based framework with LORA fine-tuning on existing natural language-LLM, achieving comparable or even better performance compared to current single-cell foundation models. Within this framework, CellHermes functions as an encoder, a predictor and an explainer for supporting a range of downstream tasks. For encoder, CellHermes encodes biological entities with embeddings that enable accurate network discovery and cross-dataset generalization; For predictor, CellHermes unifies heterogeneous downstream tasks by translating them into natural language question-answers pairs, allowing a single model to perform multi-task prediction by instruction fine-tuning; For explainer, CellHermes explains molecular mechanisms by combining attention analysis with text-based reasoning, leveraging the interpretability and reasoning capabilities of LLMs. We evaluated the ability of CellHermes as an encoder on representing genes and cells using 5 gene-level downstream tasks and 5 diverse single-cell datasets across different tissues compared with other single cell foundation models. We also evaluated CellHermes as a predictor on BioUniBench, a new benchmark of 10 tasks across 7 databases for LLMs, including perturbation response, cell fitness estimation, gene-disease association, and cell type identification. All benchmarks suggest CellHermes can unify multiple biological tasks within one model without compromising performance. By applying CellHermes as an explainer for a melanoma patients dataset, it can uncover potential key genes on tumor-reactive T cells. Together, CellHermes establishes natural language as a unifying medium for omics, offering a foundation that, in the future, potentially enables a more integrated and interpretable research loop of biological representation, prediction, and interpretability.

bioinformatics↗

bioRxiv2025

Historic Genomes Uncover Demographic Shifts and Kinship Structures in Post-Roman Central Europe

Many European towns and villages trace their origins to Early Medieval foundations. In former Roman territories, their emergence has traditionally been linked to mass migrations from outside the Roman Empire. However, recent studies have emphasised local continuity with some individual-level mobility. We generated and analysed 248 historic genomes from Late Roman (3rd and 4th century CE) and Early Medieval (5th-8th century CE) burial sites in southern Germany, comparing them to over 2,500 contemporary and Iron Age genomes in addition to 1,344 modern-day genomes from Germany, Italy and Great-Britain. Despite small inferred Early Medieval period community sizes, genetic diversity exceeded that of modern German cities. In the Altheim graveyard, established in the 5th century by a group of Northern European descent, we inferred a demographic shift in the 6th century with the integration of newcomers with ancestry typical of a nearby Roman military camp, likely as a result of the collapse of Roman state structures. We reconstructed multigenerational pedigrees and, using a novel approach to infer ancestry of unsampled relatives, inferred immediate intermarriage between incoming and local groups, with a distinct tendency for men from former Roman background marrying women of northern descent. Burial proximity correlates strongly with kinship, in some cases spanning six generations. These communities were organized around small family units, exhibited loosely patrilineal or bilateral descent patterns, practiced reproductive monogamy, and avoided close-kin marriages. Such practices reflect broader transformations in family structures that began during the Late Roman period, were transferred to small agrarian societies in the Early Medieval period, and continued to shape European societies. By the 7th century, ongoing admixture had shaped genetic diversity patterns into those resembling Central Europe today.

genetics↗

bioRxiv2024

Seasonal and comparative evidence of adaptive gene expression in mammalian brain size plasticity

Contrasting almost all other mammalian wintering strategies, Eurasian common shrews, Sorex araneus, endure winter by shrinking their brain, skull, and most organs, only to then regrow to breeding size the following spring. How such tiny mammals achieve this unique brain size plasticity while maintaining activity through the winter remains unknown. To discover potential adaptations underlying this trait, we analyzed seasonal differential expression in the shrew hypothalamus, a brain region that both regulates metabolic homeostasis and drastically changes size, and compared hypothalamus expression across species. We discovered seasonal variation in suites of genes involved in energy homeostasis and apoptosis, shrew-specific upregulation of genes involved in the development of the hypothalamic blood brain barrier and calcium signaling, as well as overlapping seasonal and comparative gene expression divergence in genes implicated in the development and progression of human neurological and metabolic disorders, including CCDC22. With high metabolic rates and facing harsh winter conditions, Sorex araneus have evolved both adaptive and plastic mechanisms to sense and regulate their energy budget. Many of these expression changes mirrored those identified in human neurological and metabolic disease, highlighting the interactions between metabolic homeostasis, brain size plasticity, and longevity.

evolutionary biology↗

bioRxiv2024

Selection for later flowering time in an orchid through frost damage and pollinator activity

Flowering time is a key trait for plant reproductive success ensuring both overlap with pollinator activity and favorable conditions for fruit development. To quantify selection on flowering time, we individually marked 1250 plants of the Small Spider Orchid, Ophrys araneola RchB., in six populations in Northern Switzerland and surveyed them during three years. We recorded the date of first flowering, frost damage, and fruiting success of individual plants. In addition, we analyzed historical records of the orchid and its only verified pollinator, the solitary bee Andrena combinata in Northern Switzerland, to estimate potential desynchronization due to climate change. We documented strong selection for later flowering driven by frost damage, with all populations showing significant selection for later flowering in at least one year. Selection for later flowering driven by pollination (fruit set) could only be analyzed in one population due to the overall low fruit set, where it was significant in one year. The historical data from between 1970 and 2019 indicated low synchronization between orchid flowering and bee occurrence, with mean flowering three weeks earlier than the mean peak of bee occurrence, corroborating selection for later flowering through fruit set. The data also showed a significant advance of flowering time and bee-occurrence in the last decades, but to a similar degree in orchids and bees, hence without an indication of increasing desynchronization through climate change. Our study shows that selection for later flowering is mostly caused by frost damage, but also by the little synchronized flowering and pollinator activity, which is however unlikely to be a consequence of climate change in this orchid.

evolutionary biology↗

bioRxiv2024

Delineating the Effective Use of Self-Supervised Learning in Single-Cell Genomics

Self-supervised learning (SSL) has emerged as a powerful method for extracting meaningful representations from vast, unlabeled datasets, already transforming computer vision and natural language processing. Similarly, in single-cell genomics (SCG), representation learning is well-recognized for offering insights into complex biological data, even more so by the advent of early foundation model approaches. However, despite these advancements, identifying scenarios in SCG where SSL outperforms traditional supervised or unsupervised learning methods remains a nuanced challenge. Furthermore, selecting the most effective pretext tasks within the SSL framework for SCG is a critical yet unresolved question. Here, we address this gap by adapting and benchmarking SSL techniques in SCG, including masked autoencoders with multiple masking strategies and contrastive learning approaches. Trained on over 20 million cells, this study rigorously examines multiple downstream tasks, including cell type prediction, gene expression reconstruction, cross-modality prediction, and data integration. Our empirical analyses underscore the nuanced role of SSL, namely in transfer learning scenarios leveraging auxiliary data or analyzing novel datasets. Masked autoencoders excel over contrastive methods in SCG, diverging from computer vision trends. Moreover, our findings reveal notable capabilities of SSL in zero-shot cell type prediction and offer insights into its potential benefits in cross-modality prediction and data integration. In summary, we study the application of SSL in SCG, minimizing model bias through simple, fully connected networks, and benchmark SSLs utility across key representation learning scenarios.

bioinformatics↗

bioRxiv2023

The anti-tubercular callyaerins target the Mycobacterium tuberculosis-specific non-essential membrane protein Rv2113

Spread of antimicrobial resistances in the pathogen Mycobacterium tuberculosis remains a public health challenge. Thus, there is a continuous need for new therapeutic options with modes-of-action differing from current antibiotics. Previously, bioactivity-guided isolation identified the callyaerins, a class of hydrophobic cyclopeptides with an unusual (Z)-2,3-di-aminoacrylamide unit, as promising antitubercular agents. In this study, we investigated the molecular mechanisms underlying their antimycobacterial properties. Structure-activity relationship studies enabled the identification of the structural determinants relevant for their antibacterial activity. The antitubercular callyaerins are bacteriostatics selectively active against M. tuberculosis, including extensively drug-resistant (XDR) strains, with minimal cytotoxicity against human cells and a promising intracellular activity in a macrophage infection model. Via spontaneous resistance mutant screens and various chemical proteomics approaches, we showed that they act by direct targeting of the non-essential, M. tuberculosis-specific putative membrane protein Rv2113, thereby triggering a complex stress response in M. tuberculosis characterized by global downregulation of lipid biosynthesis, cell division, DNA repair and replication. Our study thus not only identifies Rv2113 as a new M. tuberculosis-specific target for antitubercular drugs, which should result in less harm of the microbiome and weaker resistance development in off-target pathogens. It furthermore demonstrates that also non-essential proteins may represent efficacious targets for antimycobacterial drugs.

microbiology↗

bioRxiv2018

A large-scale systemic RNAi screen in the red flour beetle Tribolium castaneum identifies novel genes involved in arthropod muscle development

Although muscle development has been widely studied in Drosophila melanogaster there are still many gaps in our knowledge, and it is not known to which extent this knowledge can be transferred to other insects. To help in closing these gaps we participated in a large-scale RNAi screen that used the red flour beetle, Tribolium castaneum, as a screening platform. The effects of systemic RNAi were screened upon double-stranded RNA injections into appropriate muscle-EGFP tester strains. Injections into pupae were followed by the analysis of the late embryonic/early larval muscle patterns, and injections into larvae by the analysis of the adult thoracic muscle patterns. Herein we describe the results of the first-pass screens with pupal and larval injections, which covered ~8,500 and ~5,000 genes, respectively, of a total of ~16,500 genes of the Tribolium genome. Apart from many genes known from Drosophila as regulators of muscle development, a collection of genes previously unconnected to muscle development yielded phenotypes in larval body wall and leg muscles as well as in indirect flight muscles. We then present the main candidates from the pupal injection screen that remained after being processed through a series of verification and selection steps. Further, we discuss why distinct though overlapping sets of genes are revealed by the Drosophila and Tribolium screening approaches.

developmental biology↗