Biology subjects

Herrmann, L.

Publications and source records attributed to Herrmann, L..

7 recordsLinked to original sources

bioRxiv2026

Reframing enzyme function prediction as conditional generation

Enzymes frequently exhibit promiscuous activity beyond their native roles, providing starting-points for new functions. Finding these promiscuous enzymes, especially for non-native chemical transformations, is challenging but highly valuable, as they promise novel, sustainable solutions for chemistry and biotechnology. However, current machine learning methods are poorly suited to discovering unseen chemistry as they often frame function prediction as closed set classification or a retrieval task. Here, we present Fluxion, a generative deep learning framework that learns enzymatic catalysis by modeling dynamic electron flow trajectories across the enzyme's catalytic residues. By combining both synthetic chemistry and biochemical datasets with protein language model representations, Fluxion generates multi-step electron-flow trajectories analogous to the arrow-pushing representations used to describe enzyme reaction mechanisms. Generation is conditioned on enzyme context, including the enzyme sequence, catalytic residues, substrates, and cofactors. We show that this conditioning allows Fluxion to learn enzyme-dependent regioselectivity across cytochrome P450 enzymes with different sequences shifting the predicted reaction sites for the same substrate. We then demonstrate that Fluxion's embeddings are useful for downstream tasks, such as specificity prediction on two experimental datasets, with and without finetuning. Finally, we show that Fluxion has the potential to transfer synthetic chemical logic to biology; it can generate the observed non-native product from real-world non-native directed evolution screens. Our results establish a proof of concept that generative modeling through mechanistic representations of enzymes can shift enzyme function prediction beyond static database retrieval and closed set classification to function generation. This conceptual framework provides a stepping stone towards an in silico generative method to discover non-native biocatalysts.

biochemistry↗

bioRxiv2025

Biologically Guided Variational Inference for Interpretable Multimodal Single-Cell Integration and Mechanistic Discovery

Multi-omics technologies allow detailed characterization of cell types and states across omics layers as well as chemical and genetic perturbations. Variational autoencoders have become a cornerstone of single-cell data integration; however, they are often implemented as black-box models, requiring post hoc interpretation using known markers, pathways, or regulators to make sense of their latent representations. NetworkVI fundamentally flips this paradigm by incorporating biological knowledge directly into the model architecture. By embedding co-regulation networks derived from topologically associated domains and structured ontologies such as the Gene Ontology (GO), NetworkVI introduces a biologically-informed inductive bias that promotes the preservation of meaningful variation during integration while enforcing interpretability at both the gene and GO levels. NetworkVI achieves state-of-the-art data integration, modality imputation, and cell label transfer across bimodal and trimodal datasets. Beyond integration, here we show that NetworkVI facilitates ontology-guided hypothesis generation by exploiting established associations between genes, structured cellular programs, and regulatory domains to interpretably model cellular identities. Furthermore, decomposition of GO activation spaces resolves lineage-specific functional states within immune cell types, including quiescent, inflammatory, and transitional monocyte subpopulations, that are invisible to transcriptomic clustering. NetworkVI prioritizes GO-term programs associated with immunosenescence, consistent with age-associated immune dysregulation and reveals candidate immune evasion mechanisms consistent with CD58 loss in a Perturb-CITE-seq melanoma dataset.

bioinformatics↗

bioRxiv2024

How to Predict Effective Drug Combinations - Moving beyond Synergy Scores

To improve our understanding of multi-drug therapies, cancer cell line panels screened with drug combinations are frequently studied using machine learning (ML). ML models trained on such data typically focus on predicting synergy scores, which support drug development and repurposing efforts but have limitations when deriving personalized treatment recommendations. To simulate a more realistic personalized treatment scenario, we pioneer ML models that predict the relative growth inhibition (instead of synergy scores), and that can be applied to previously unseen cell lines. Our approach is highly flexible: it enables the reconstruction of dose-response curves and matrices, as well as various measures of drug sensitivity (and synergy) from model predictions, which can finally even be used to derive cell line-specific prioritizations of both mono- and combination therapies.

bioinformatics↗

bioRxiv2024

Functional interaction between transcription factor Sfp1 and the NuA4 complex in response to nutrient availability

Ribosome biogenesis is a crucial process requiring enormous transcriptional output. In budding yeast, the expression of 138 ribosomal protein (RP) genes and over 200 ribosome biogenesis (RiBi) genes is regulated by an intricate network of factors, including the nutrient-sensitive transcription activator Sfp1 and the NuA4 coactivator/acetyltransferase complex. Nutrient starvation or inhibition of TORC1 by rapamycin leads to repression of RP and RiBi genes, in part through blocking Sfp1 nuclear localization and NuA4-dependent chromatin acetylation. Here, we demonstrate that Sfp1 physically interacts with NuA4 in a TORC1-dependent manner. Our results indicate that Sfp1, along with NuA4, regulate the transcription of RiBi and RP genes via distinct mechanisms depending on promoter architectures. Sfp1 promotes histone acetylation at the promoters without affecting NuA4 recruitment. In contrast, NuA4 does impact Sfp1 binding but specifically at two classes of RP genes. Importantly, NuA4 acetylates Sfp1 at lysines 655 and 657, regulating its function. Cells expressing Sfp1 with acetyl-mimicking mutations exhibit increased expression of RiBi genes while RP genes remain stable. However, the same mutants lead to the loss of Sfp1 binding/activity at RiBi genes when cells are under non-optimal growth conditions. Mimicking constitutive acetylation of Sfp1 also limits the transcriptional burst of RP genes upon addition of glucose. Altogether, these results draw an intricate functional relationship between Sfp1 and NuA4 to control ribosome biogenesis, fine-tuning transcription output in different growth conditions.

molecular biology↗

bioRxiv2024

Characterization of antimalarial activity of artemisinin-based hybrid drugs

In response to the spread of artemisinin (ART) resistance, ART-based hybrid drugs were developed and their activity profile was characterized against drug-sensitive and drug-resistant Plasmodium falciparum parasites. Two hybrids were found to display parasite growth reduction, stage-specificity, speed of activity, additivity of activity in drug combinations, and stability in hepatic microsomes of similar levels to those displayed by dihydroartemisinin (DHA). Conversely, the rate of chemical homolysis of the peroxide bonds is slower in the hybrids than in DHA. From a mechanistic perspective, heme plays a central role in the chemical homolysis of peroxide and in inhibiting heme detoxification and disrupting parasite heme redox homeostasis. The hybrid exhibiting slow homolysis of peroxide bonds was more potent in reducing the viability of ART-resistant parasites in a ring-stage survival assay than the hybrid exhibiting fast homolysis. However, both hybrids showed some limited activity against ART-induced quiescent parasites in the quiescent-stage survival assay. Our findings are consistent with previous results showing that slow homolysis of peroxide-containing drugs may retain activity against proliferating ART-resistant parasites. However, our data suggest that this property does not overcome the limited activity of peroxides in killing non-proliferating parasites in a quiescent state. GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=131 SRC="FIGDIR/small/577447v1_ufig1.gif" ALT="Figure 1"> View larger version (27K): org.highwire.dtl.DTLVardef@6ea6b6org.highwire.dtl.DTLVardef@831dedorg.highwire.dtl.DTLVardef@21ee5dorg.highwire.dtl.DTLVardef@1a842b7_HPS_FORMAT_FIGEXP M_FIG C_FIG Hepatic and cell-host-mediated metabolism are responsible for short plasma half-lives of antimalarial artemisinins (ARTs), illustrated here by dihydroartemisinin (DHA). ART-based hybrid drugs that overcome rapid degradation can facilitate activity against ART-resistant parasites, as illustrated by hybrid 1.

microbiology↗

bioRxiv2023

Interplay of RAP2 GTPase and the cytoskeleton in Hippo pathway regulation

The Hippo signaling is instrumental in regulating organ size, regeneration, and carcinogenesis. The cytoskeleton emerges as a primary Hippo signaling modulator. Its structural alterations in response to environmental and intrinsic stimuli control Hippo kinase cascade activity. However, the precise mechanisms underlying the cytoskeleton regulation of Hippo signaling are not fully understood. RAP2 GTPase is known to mediate the mechanoresponses of Hippo signaling via activating the core Hippo kinases LATS1/2 through MAP4Ks and MST1/2. Here we show the pivotal role of the reciprocal regulation between RAP2 GTPase and the cytoskeleton in Hippo signaling. RAP2 deletion undermines the responses of the Hippo pathway to external cues tied to RhoA GTPase inhibition and actin cytoskeleton remodeling, such as energy stress and serum deprivtion. Notably, RhoA inhibitors and actin disruptors fail to activate LATS1/2 effectively in RAP2-deficient cells. RNA sequencing highlighted differential regulation of both actin and microtubule networks by RAP2 gene deletion. Consistently, Taxol, a microtubule-stabilizing agent, was less effective in activating LATS1/2 and inhibiting cell growth in RAP2 and MAP4K4/6/7 knockout cells. In summary, our findings position RAP2 as a central integrator of cytoskeletal signals for Hippo signaling, which offers new avenues for understanding Hippo regulation and therapeutic interventions in Hippo-impaired cancers.

cell biology↗

bioRxiv2019

The anterior insula channels prefrontal expectancy signals during affective processing

Expectancy shapes our perception of impending events. Although such an interplay between cognitive and affective processes is often impaired in mental disorders, it is not well understood how top-down expectancy signals modulate future affect. We therefore track the information flow in the brain during cognitive and affective processing segregated in time using task-specific cross-correlations. Participants in two independent fMRI studies (N1 = 37 & N2 = 55) were instructed to imagine a situation with affective content as indicated by a cue, which was then followed by an emotional picture congruent with expectancy. To correct for intrinsic covariance of brain function, we calculate resting-state cross-correlations analogous to the task. First, using factorial modeling of delta cross-correlations (task-rest) of the first study, we find that the magnitude of expectancy signals in the anterior insula cortex (AIC) modulates the BOLD response to emotional pictures in the anterior cingulate and dorsomedial prefrontal cortex in opposite directions. Second, using hierarchical linear modeling of lagged connectivity, we demonstrate that expectancy signals in the AIC indeed foreshadow this opposing pattern in the prefrontal cortex. Third, we replicate the results in the second study using a higher temporal resolution, showing that our task-specific cross-correlation approach robustly uncovers the dynamics of information flow. We conclude that the AIC arbitrates the recruitment of distinct prefrontal networks during cued picture processing according to triggered expectations. Taken together, our study provides new insights into neuronal pathways channeling cognition and affect within well-defined brain networks. Better understanding of such dynamics could lead to new applications tracking aberrant information processing in mental disorders.

neuroscience↗