MAMichael F. Adamer

Michael F. Adamer

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Michael F. Adamer has a total of 33 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Bacterial protein function prediction via multimodal deep learning

    by Lukas Miklautz, Leyden Fernandez, Giulia Muzio, Michael Adamer, Karsten Borgwardt, Kemal Avican

    Bacterial proteins are specialized with extensive functional diversity for survival in diverse and stressful environments. A significant portion of these proteins remains functionally uncharacterized, limiting our understanding of bacterial survival mechanisms. Hence, we develop Deep Expression STructure (DeepEST), a multimodal deep learning framework that predicts protein function in bacteria by assigning Gene Ontology (GO) terms. DeepEST comprises two modules: a multi-layer perceptron that uses gene expression and gene location, and a protein structure-based module that uses predicted structures and amino acid sequences. Within DeepEST, we integrate these modules through a learnable weighted linear combination and introduce a novel masked loss function to fine-tune the structure-based predictor for bacterial species. These choices are well suited for bacteria, in which functionally related genes frequently colocalize and are cotranscribed within operons, allowing transcription dynamics to serve as informative, condition-dependent signals. We show that DeepEST outperforms existing protein function prediction methods on a 25-species benchmark, relying solely on amino acid sequence or protein structure. Averaged over the 25 species, DeepEST reaches a protein-centric F max of 0.689 and a term-centric micro-AUPRC of 0.756, a relative improvement of 4.4% and 11.6% over the second-best method, the structure-based predictor DeepFRI. Moreover, DeepEST predicts GO terms for unclassified hypothetical proteins across 25 human bacterial pathogens, supporting their experimental characterization. DeepEST bridges organism-specific data integration and structure-based transfer learning, providing a method tailored for bacterial protein function prediction in settings with structural and multi-condition expression data.

  • Open Access

    Bacterial protein function prediction via multimodal deep learning

    by Giulia Muzio, Michael Adamer, Leyden Fernandez, Lukas Miklautz, Karsten Borgwardt, Kemal Avican

    Bacterial proteins are specialized with extensive functional diversity for survival in diverse and stressful environments. A significant portion of these proteins remains functionally uncharacterized, limiting our understanding of bacterial survival mechanisms. Hence, we developed Deep Expression STructure (DeepEST), a multimodal deep learning framework designed to accurately predict protein function in bacteria by assigning Gene Ontology (GO) terms. DeepEST comprises two modules: a multi-layer perceptron that takes gene expression and gene location as input features, and a protein structure-based predictor. Within DeepEST, we integrated these modules through a learnable weighted linear combination and introduced a novel masked loss function to fine-tune the structure-based predictor for bacterial species. These modeling choices are particularly well suited for bacteria due to the spatial organization of their circular genomes. Functionally related genes frequently co-localize and are co-transcribed within operons, allowing transcription dynamics to serve as crucial, condition-dependent regulatory signals. We show that DeepEST outperforms existing protein function prediction methods on a 25-species benchmark, relying solely on amino acid sequence or protein structure. Moreover, DeepEST predicts GO terms for unclassified hypothetical proteins across 25 human bacterial pathogens, facilitating the design of experimental setups for characterization studies. By combining expression, localization, and structure information in a unified deep learning framework, DeepEST bridges organism-specific data integration and structure-based transfer learning, providing a method tailored for bacterial protein function prediction in settings with structural and multi-condition expression data. Availability: Accompanying code and data can be found under https://github.com/BorgwardtLab/DeepEST .

  • Open Access

    The magnitude vector of images

    by Michael F. Adamer, Edward De Brouwer, Leslie O’bRay, Bastian Rieck

    The magnitude of a finite metric space has recently emerged as a novel invariant quantity, allowing to measure the effective size of a metric space. Despite encouraging first results demonstrating the descriptive abilities of the magnitude, such as being able to detect the boundary of a metric space, the potential use cases of magnitude remain under-explored. In this work, we investigate the properties of the magnitude on images, an important data modality in many machine learning applications. By endowing each individual image with its own metric space, we are able to define the concept of magnitude on images and analyse the individual contribution of each pixel with the magnitude vector. In particular, we theoretically show that the previously known properties of boundary detection translate to edge detection abilities in images. Furthermore, we demonstrate practical use cases of magnitude for machine learning applications and propose a novel magnitude model that consists of a computationally efficient magnitude computation and a learnable metric. By doing so, we address one computational hurdle that used to make magnitude impractical for many applications and open the way for the adoption of magnitude in machine learning research.


  • Open Access

    Biomarker identification by interpretable maximum mean discrepancy

    by Michael F Adamer, Sarah C Brüningk, Dexiong Chen, Karsten Borgwardt

    In many biomedical applications, we are confronted with paired groups of samples, such as treated versus control. The aim is to detect discriminating features, i.e. biomarkers, based on high-dimensional (omics-) data. This problem can be phrased more generally as a two-sample problem requiring statistical significance testing to establish differences, and interpretations to identify distinguishing features. The multivariate maximum mean discrepancy (MMD) test quantifies group-level differences, whereas statistically significantly associated features are usually found by univariate feature selection. Currently, few general-purpose methods simultaneously perform multivariate feature selection and two-sample testing. We introduce a sparse, interpretable, and optimized MMD test (SpInOpt-MMD) that enables two-sample testing and feature selection in the same experiment. SpInOpt-MMD is a versatile method and we demonstrate its application to a variety of synthetic and real-world data types including images, gene expression measurements, and text data. SpInOpt-MMD is effective in identifying relevant features in small sample sizes and outperforms other feature selection methods such as SHapley Additive exPlanations and univariate association analysis in several experiments. The code and links to our public data are available at https://github.com/BorgwardtLab/spinoptmmd.


  • Open Access

    SIMBSIG: similarity search and clustering for biobank-scale data

    by Michael F Adamer, Eljas Roellin, Lucie Bourguignon, Karsten Borgwardt

    Summary: In many modern bioinformatics applications, such as statistical genetics, or single-cell analysis, one frequently encounters datasets which are orders of magnitude too large for conventional in-memory analysis. To tackle this challenge, we introduce SIMBSIG (SIMmilarity Batched Search Integrated GPU), a highly scalable Python package which provides a scikit-learn-like interface for out-of-core, GPU-enabled similarity searches, principal component analysis and clustering. Due to the PyTorch backend, it is highly modular and particularly tailored to many data types with a particular focus on biobank data analysis. Availability and implementation: SIMBSIG is freely available from PyPI and its source code and documentation can be found on GitHub (https://github.com/BorgwardtLab/simbsig) under a BSD-3 license.

  • Open Access

    Tuning of cytokine signaling through imbalanced abundances of receptors and kinases

    by Léa Sta, Guillaume Voisinne, Jesse Cotari, Michael F. Adamer, Carmen Molina-París, Grégoire Altan-Bonnet

    Cells use surface receptors to detect extra-cellular ligands ( e . g ., growth factors or cytokines) and engender signaling cascades. We explore the effect of cell-to-cell variability in receptor subunit copy number for cells responding to γc cytokines ( e . g ., IL-2, IL-7 and IL-15). We find that primary T cells expressing higher levels of the common γ c receptor chain have weaker responses to IL-7, both in terms of lowered STAT5 phosphorylation amplitude and higher EC 50 . A mathematical model that accounts for abundance imbalance ( e . g ., insufficient expression of JAK kinases compared to the number of receptors, or γ c competition for other receptor subunit chains) predicts the formation of non-signalling complexes, consistent with the observed cellular behaviour. This type of built-in limit on signaling responses illustrates how phenotypic heterogeneity generates biological functional diversity. Significance Statement: Cells rely on cytokines to coordinate their activation, differentiation, proliferation and survival. In particular, γ c cytokines (interleukins IL-2, 4, 7, 9, 15, and 21) regulate the fate of leukocytes. The signaling cascade induced by these cytokines is relatively simple, and involves the phosphorylation of receptor-associated Janus-like kinases (JAK). Here, we explore the cell-to-cell variability of cytokine responses in primary mouse T cells, and find a paradoxical and quantitative imprint of receptor expression levels and other signaling components. For instance, high abundance of the common γ c chain reduces cytokine responses (both in terms of signaling amplitudes and characteristic cytokine concentrations triggering 50% of the response). We develop mathematical models to quantify how limited abundances of signaling components ( e . g ., JAK or other cytokine receptor subunit chains) may explain our experimental observations. We conclude by generalizing this observation of cell-to-cell signaling variability to other ligand-receptor-kinase systems.


  • Open Access

    SIMBSIG: Similarity search and clustering for biobank-scale data

    by Michael F. Adamer, Eljas Roellin, Lucie Bourguignon, Karsten Borgwardt

    Summary In many modern bioinformatics applications, such as statistical genetics, or single-cell analysis, one frequently encounters datasets which are orders of magnitude too large for conventional in-memory analysis. To tackle this challenge, we introduce SIMBSIG, a highly scalable Python package which provides a scikit-learn-like interface for out-of-core, GPU-enabled similarity searches, principal component analysis, and clustering. Due to the PyTorch backend it is highly modular and particularly tailored to many data types with a particular focus on biobank data analysis. Availability SIMBSIG is freely available from PyPI and its source code and documentation can be found on GitHub (https://github.com/BorgwardtLab/simbsig) under a BSD-3 license. Contact michael.adamer{at}bsse.ethz.ch

  • Open Access

    reComBat: batch-effect removal in large-scale multi-source gene-expression data integration

    by Michael F Adamer, Sarah C Brüningk, Alejandro Tejada-Arranz, Fabienne Estermann, Marek Basler, Karsten Borgwardt

    With the steadily increasing abundance of omics data produced all over the world under vastly different experimental conditions residing in public databases, a crucial step in many data-driven bioinformatics applications is that of data integration. The challenge of batch-effect removal for entire databases lies in the large number of batches and biological variation which can result in design matrix singularity. This problem can currently not be solved satisfactorily by any common batch-correction algorithm. We present reComBat, a regularized version of the empirical Bayes method to overcome this limitation and benchmark it against popular approaches for the harmonization of public gene expression data (both microarray and bulkRNAsq) of the human opportunistic pathogen Pseudomonas aeruginosa. Batch-effects are successfully mitigated while biologically meaningful gene expression variation is retained. reComBat fills the gap in batch-correction approaches applicable to large-scale, public omics databases and opens up new avenues for data-driven analysis of complex biological processes beyond the scope of a single study. The code is available at https://github.com/BorgwardtLab/reComBat, all data and evaluation code can be found at https://github.com/BorgwardtLab/batchCorrectionPublicData


  • Open Access

    reComBat: batch-effect removal in large-scale multi-source gene-expression data integration

    by Michael F. Adamer, Sarah C. Brüningk, Alejandro Tejada-Arranz, Fabienne Estermann, Marek Basler, Karsten M. Borgwardt

    Motivation: With the steadily increasing abundance of omics data produced all over the world under vastly different experimental conditions residing in public databases, a crucial step in many data-driven bioinformatics applications is that of data integration. The challenge of batch-effect removal for entire databases lies in the large number of batches and biological variation which can result in design matrix singularity. This problem can currently not be solved satisfactorily by any common batch-correction algorithm. Results: We present reComBat , a regularized version of the empirical Bayes method to overcome this limitation and benchmark it against popular approaches for the harmonization of public gene expression data (both microarray and bulkRNAsq) of the human opportunistic pathogen Pseudomonas aeruginosa . Batch-effects are successfully mitigated while biologically meaningful gene expression variation is retained. reComBat fills the gap in batch-correction approaches applicable to large-scale, public omics databases and opens up new avenues for data-driven analysis of complex biological processes beyond the scope of a single study. Contact: michael.adamer@bsse.ethz.ch Availability: The code is available at https://github.com/BorgwardtLab/reComBat , all data and evaluation code can be found at https://github.com/BorgwardtLab/batchCorrectionPublicData


  • Open Access

    Families of toric chemical reaction networks

    by Michael F. Adamer, Martin Helmer

    We study families of chemical reaction networks whose positive steady states are toric, and therefore can be parameterized by monomials. Families are constructed algorithmically from a core network; we show that if a family member is multistationary, then so are all subsequent networks in the family. Further, we address the questions of model selection and experimental design for families by investigating the algebraic dependencies of the chemical concentrations using matroids. Given a family with toric steady states and a constant number of conservation relations, we construct a matroid that encodes important information regarding the steady state behaviour of the entire family. Among other things, this gives necessary conditions for the distinguishability of families of reaction networks with respect to a data set of measured chemical concentrations. We illustrate our results using multi-site phosphorylation networks.