Aleksandra M. Walczak

Computational identification of responding T cell clones from longitudinal repertoire sequencing

This paper introduces CloneSearch, a method to identify expanding and contracting T cell clones from longitudinal TCR sequencing without requiring antigen-specific information.

Paraplume: A fast and accurate paratope prediction method provides insights into repertoire-scale binding dynamics

Paraplume predicts paratopes from antibody sequences using protein language models, improving downstream tasks and revealing evolutionary dynamics in antibody repertoires.

Modeling Protein Evolution via Generative Inference From Monte Carlo Chains to Population Genetics

This study compares three simulation schemes for modeling protein evolution, finding that population genetics dynamics best capture phylogenetic correlations and selective sweeps without requiring additional data.

Noradrenaline prolongs the state of motor arrest via cerebrospinal fluid signaling

This study shows noradrenaline prolongs motor arrest through cerebrospinal fluid signaling via Adra1a receptors on ependymal radial glia.

Dynamics of memory B cells and plasmablasts in healthy individuals

The paper infers memory B cell and plasmablast dynamics using statistical methods, finding slow clone degradation and high plasmablast differentiation rates.

Single cells can resolve graded stimuli

The paper quantifies information transduced through the ERK pathway in single cells and populations, showing single cells can resolve graded stimuli with over 2 bits of information.

Coarse-graining reveals collective predictive information in a sensory population

The paper identifies collective variables in retinal ganglion cells that best predict future neural activity using an information bottleneck method.

Profiling a large HIV-1 elite neutralizer cohort reveals remarkable CD4bs bNAb for HIV-1 prevention and therapy

The paper identifies a new broadly neutralizing antibody, 04_A06, with high potency and breadth against HIV-1 strains, showing potential for prevention and therapy.

Learning evolutionary parameters from genealogies using allelic trees

The paper introduces CBA, a probabilistic method to infer genetic and growth rates from cell populations, using a backbone tree and monogram to model evolutionary processes.

Optimizing information transmission in neural induction constrains cell surface contacts of ascidian embryos

The paper investigates how optimizing information transmission between FGF and ERK signaling in ascidian embryos depends on cell surface contact geometry, finding that optimal contacts align with experimental observations.

Inferring resource competition in microbial communities from time series

This study uses spectral methods to infer resource competition in microbial communities from time series data, showing they outperform simple correlations.

TULIP - a Transformer based Unsupervised Language model for Interacting Peptides and T-cell receptors that generalizes to unseen epitopes

TULIP is an unsupervised transformer model for predicting TCR epitope binding, outperforming existing methods on unseen epitopes by addressing data limitations and bias.

Learning the differences: a transfer-learning approach to predict antigen immunogenicity and T-cell receptor specificity

This paper introduces diffRBM, a transfer-learning model using Restricted Boltzmann Machines to predict antigen immunogenicity and T-cell receptor specificity, achieving performance comparable to existing methods.

Granger causality analysis for calcium transients in neuronal networks: challenges and improvements

This paper improves Granger causality analysis for calcium imaging data to better infer information flow in neuronal networks.

Benchmarking solutions to the T-cell receptor epitope prediction problem: IMMREP22 workshop report

The paper benchmarks 23 TCR-epitope prediction models, finding that using paired-chain alpha-beta and CDR1/2/V/J data improves performance, and simple distance-based methods can be competitive.

Synthetic reconstruction of the hunchback promoter specifies the role of Bicoid, Zelda and Hunchback in the dynamics of its transcription

The paper shows that the Bicoid-hunchback system's transcription dynamics can be explained by an equilibrium model with Bicoid as the sole positional information source, with Zelda and Hunchback modulating the process.

Probing T-cell response by sequence-based probabilistic modeling

This paper develops a probabilistic machine learning model to identify antigen-specific TCRs using sequence data and clone abundance, enabling personalized discrimination of specific versus non-specific T-cell responses.

Immune Fingerprinting through Repertoire Similarity

This study demonstrates that immune repertoires can uniquely identify individuals with high accuracy using a classifier based on repertoire similarity.

Flexible machine learning prediction of antigen presentation for rare and common HLA-I alleles

This paper introduces a flexible machine learning method for predicting antigen presentation by HLA-I alleles and shows its improved accuracy for rare alleles.

Constraints and limitations on decoding positional information: the Bicoid case-study

The paper examines the challenges in quantitatively modeling transcriptional regulation by the Bicoid gradient, emphasizing the role of bursty transcription dynamics.

OLGA: fast computation of generation probabilities of B- and T-cell receptor amino acid sequences and motifs

OLGA calculates amino acid sequence generation probabilities for B- and T-cell receptors using dynamic programming, showing agreement with published data and potential for vaccine design.

Active degradation of a regulator controls coordination of downstream genes

The paper investigates how active degradation of the MarA regulator affects coordination of downstream genes in E. coli, finding trade-offs between coordination speed and maximum coordination.

3 minutes to precisely measure morphogen concentration

The paper measures Bicoid morphogen concentration dynamics in fruit fly embryos using a modified MS2-MCP reporter and shows that a 750 bp promoter region can recapitulate endogenous expression with high precision within 3 minutes.

Physical epistatic landscape of antibody binding affinity

The paper maps the epistatic landscape of antibody binding affinity using a null model and Tite-Seq data, showing that epistasis enhances the mutational accessibility of high-affinity antibodies.

Disorder and the neural representation of complex odors: smelling in the real world

The paper proposes a theory of olfactory circuit architecture in Drosophila that uses disorder to process complex odors, tested with empirical data.

Population dynamics of immune repertoires

This study uses mathematical models informed by experiments to show that memory immune repertoire evolution is non-neutral, driven by pathogenic history, while naive repertoires follow neutral dynamics.

Habitat Fluctuations Drive Species Covariation in the Human Microbiota

The paper shows that habitat fluctuations explain correlations in human microbiota species abundances and identifies key metabolic pathways involved.