Biology subjects

Haag, J.

Publications and source records attributed to Haag, J..

15 recordsLinked to original sources

bioRxiv2026

Time-Series Foundation Models for Cognitive Workload Classification using Eye-Tracking Data

Cognitive workload (CWL) assessment is relevant to a range of applications, such as monitoring driver fatigue and pilot attention, surgeon workload during complex procedures, and astronaut cognitive fatigue during long-duration missions. Eye-tracking datasets are generally small, which hinders the generalizability of the AI models. Time-series foundation models (TSFMs) have shown promise in mitigating this limitation as they are pretrained on large corpora and can be effectively fine-tuned with limited task-specific data. Although cognitive workload measurement has been explored in controlled settings, the ability of TSFMs to generalize to unseen individuals for CWL classification from eye-tracking data has not been studied. In this paper, two TSFMs, MOMENT and Moirai, were evaluated for CWL classification and compared against CNN, FFN, and LSTM baselines on two publicly available eye-tracking datasets. We used subject-level five-fold cross-validation in which data from each test subject were held out during training. We report accuracy, AUC, and F1-score with 95% confidence intervals. On the class-balanced EM-COGLOAD dataset, the pretrained TSFMs generalized markedly better to unseen subjects, with Moirai reaching 0.918 AUC and MOMENT 0.882 AUC, while the task-specific baselines remained near 0.70 AUC. On the class-imbalanced COLET dataset, even though the performance of all the models decreased, MOMENT was the most robust (0.670 AUC). Both TSFMs were also the most stable across held-out subjects, indicating that pretrained representations generalize more consistently across individuals than task-specific models in data-scarce eye-tracking settings.

bioinformatics↗

bioRxiv2026

Quantifying and Predicting the Difficulty of Multiple Sequence Alignment with AlDiScore

Multiple Sequence Alignment (MSA) constitutes an important and frequent operation in molecular sequence data analysis. There exist numerous tools, algorithms, and criteria to infer an MSA. This plethora of available approaches to MSA may induced an ensemble of divergent MSAs for the same underlying unaligned sequence set. Even a single MSA tool may infer distinct MSAs when varying the input parameters. Hence, when using a diversified set of MSA algorithms and parameterizations, the observed dispersion within an MSA ensemble expresses the difficulty of inferring a robust alignment. We refer to this notion as MSA difficulty. As downstream analyses heavily rely on the MSA, characterizing MSA difficulty for a given unaligned sequence set is critical. Initially, we show that measures of dispersion within diversified MSA ensembles can reliably predict MSA difficulty. We then assess the adequacy of these measures by computing the average reference-based distance between the MSAs in the MSA ensemble and its corresponding structural reference MSA and subsequently comparing this distance to the corresponding reference-free average distance over all MSA pairs in the ensemble. We find that Blackburne and Whelans dpos alignment metric is most appropriate as its reference-free [Formula] counterpart most accurately approximates the reference-based difficulty computed on BAliBASE reference data. We therefore use [Formula] to quantify MSA difficulty on a scale from 0 (easy) to 1 (difficult). Next, we introduce the AlDiScore open-source tool, which uses machine learning to directly and reliably predict reference-free difficulty scores from unaligned sequence sets to completely omit expensive MSA computations. The underlying regression model relies upon a large set of features, including sampling-based measures of transitive consistency. We trained our AlDiScore model on a diverse collection of empirical datasets from BAliBASE, TreeBASE, and published studies. Subsequently, we demonstrate that AlDiScore attains an R2 of 0.89 and of 0.84 on unseen AA and DNA sequence sets extracted from the PANDIT v17 database. Finally, we show that there is no correlation between MSA difficulty and the corresponding phylogenetic difficulty of the respective MSA.

bioinformatics↗

bioRxiv2026

Motion-based depth estimation in Drosophila

Inferring three-dimensional structure from two-dimensional visual input requires a second perspective, either in time or in space. In flying animals, the first of these options lends itself as an ideal source for constructing internal representations of the environment, since their continuous self-motion automatically produces an optic flow. How neural circuits transform this signal into estimates of object distance, and what form such representations take, remains unknown. We address this problem in the visual system of the fruit fly. By immersing flies in a virtual environment while recording population-level calcium activity of motion-sensitive neurons, we extract an optic flow-based map of space. This map contains accurate estimates of object distance computed through spatial integration of elementary motion detector signals. Using electrophysiological recordings, we identify postsynaptic wide-field neurons as a cellular substrate for this operation. In behavioural experiments, we show that motion vision is essential for depth perception; without access to visual motion information, the free flight trajectories of motion-blind flies inevitably end in collisions. Our experiments link population-level neural activity to behaviourally relevant representations of environmental structure and demonstrate that motion vision is essential for navigation in three-dimensional space. One Sentence SummaryPopulation activity of motion-sensitive neurons accurately captures object distance and is required for navigation in three dimensions.

neuroscience↗

bioRxiv2025

Pythia 2.0: New Data, New Prediction Model, New Features

Maximum Likelihood (ML) based phylogenetic inference is time- and resource-intensive, especially when initiating multiple independent inferences from distinct comprehensive tree topologies. Performing multiple independent inferences is often required to (sufficiently) explore the vast search space of possible unrooted binary tree topologies. Yet, these independent inferences do not necessarily converge to a single phylogeny or at least topologically highly similar trees. While for easy-to-analyze multiple sequence alignments (MSAs), one is likely to obtain a conclusive, single phylogeny, difficult-to-analyze MSAs yield topologically highly distinct, yet statistically indistinguishable tree topologies. In 2022, we proposed a compute-intensive approach to quantify the inherent difficulty of a phylogenetic analysis for a specific, given MSA, and also trained a machine-learning based prediction model called Pythia to substantially reduce the computational cost of determining the difficulty. Pythia can predict the difficulty for a given MSA with high accuracy, while being substantially faster than even a single ML tree inference. Pythia predicts the difficulty on a scale from 0 (easy) to 1 (difficult). Here, we present all improvements to Pythia that we have introduced since our initial publication in 2022. We trained a new prediction model using approximately three times more MSAs and a new type of machine learning model. We improved the runtime of two feature computations, and we also introduced two additional prediction features. Our latest version Pythia 2.0 is slightly more accurate than our initial version and is also approximately twice as fast. Finally, we also present and make available, the novel and easy-to-use command line tool PyDLG that allows to compute the ground-truth difficulty seamlessly for a given MSA. This ground-truth difficulty can be used, for instance, as a prediction target for training a new Pythia model.

bioinformatics↗

bioRxiv2024

Columnar cholinergic neurotransmission onto T5 cells of Drosophila

Several nicotinic and muscarinic acetylcholine receptors (AChRs) are expressed in the brain of Drosophila melanogaster. However, the contribution of different AChRs to visual information processing remains poorly understood. T5 cells are the primary motion-sensing neurons in the OFF pathway and receive input from four different columnar cholinergic neurons, Tm1, Tm2, Tm4 and Tm9. We reasoned that different AChRs in T5 postsynaptic sites might contribute to direction selectivity, a central feature of motion detection. We show that the nicotinic nAChR1, nAChR4, nAChR5 and nAChR7 subunits localize on T5 dendrites. By targeting synaptic markers specifically to each cholinergic input neuron, we find a prevalence of the nAChR5 in Tm1-, Tm2- and Tm4-to-T5 synapses and of nAChR7 in Tm9-to-T5 synapses. Knock-down of nAChR4, nAChR5, nAChR7, or mAChR-B individually in T5 cells alters the optomotor response and reduces T5 directional selectivity. Our findings indicate a differential contribution of postsynaptic receptors to input visual processing and, thus, to the computation of motion direction in T5 cells.

neuroscience↗

bioRxiv2024

Pandora: A Tool to Estimate Dimensionality Reduction Stability of Genotype Data

MotivationGenotype datasets typically contain a large number of single nucleotide polymorphisms for a comparatively small number of individuals. To identify similarities between individuals and to infer an individuals origin or membership to a population, dimensionality reduction techniques are routinely deployed. However, inherent (technical) difficulties such as missing or noisy data need to be accounted for when analyzing a lower dimensional representation of genotype data, and the intrinsic uncertainty of such analyses should be reported in all studies. However, to date, there exists no stability assessment technique for genotype data that can estimate this uncertainty. ResultHere, we present Pandora, a stability estimation framework for genotype data based on bootstrapping. Pandora computes an overall score to quantify the stability of the entire embedding, infers per-individual support values, and also deploys a k-means clustering approach to assess the uncertainty of assignments to potential cultural groups. Using published empirical and simulated datasets, we demonstrate the usage and utility of Pandora for studies that rely on dimensionality reduction techniques. Availability and ImplementationPandora is available on GitHub https://github.com/tschuelia/Pandora. Contactjulia.haag@h-its.org Supplementary informationSupplementary data are available online.

bioinformatics↗

bioRxiv2024

Predicting Phylogenetic Bootstrap Values viaMachine Learning

SummaryEstimating the statistical robustness of the inferred tree(s) constitutes an integral part of most phylogenetic analyses. Commonly, one computes and assigns a branch support value to each inner branch of the inferred phylogeny. The most widely used method for calculating branch support on trees inferred under Maximum Likelihood (ML) is the Standard, non-parametric Felsenstein Bootstrap Support (SBS). Due to the high computational cost of the SBS, a plethora of methods has been developed to approximate it, for instance, via the Rapid Bootstrap (RB) algorithm. There have also been attempts to devise faster, alternative support measures, such as the SH-aLRT (Shimodaira-Hasegawalike approximate Likelihood Ratio Test) or the UltraFast Bootstrap 2 (UFBoot2) method. Those faster alternatives exhibit some limitations, such as the need to assess model violations (UFBoot2) or meaningless low branch support intervals (SH-aLRT). Here, we present the Educated Bootstrap Guesser (EBG), a machine learning-based tool that predicts SBS branch support values for a given input phylogeny. EBG is on average 9.4 ({sigma} = 5.5) times faster than UFBoot2. EBG-based SBS estimates exhibit a median absolute error of 5 when predicting SBS values between 0 and 100. Furthermore, EBG also provides uncertainty measures for all per-branch SBS predictions and thereby allows for a more rigorous and careful interpretation. EBG can predict SBS support values on a phylogeny comprising 1654 SARS-CoV2 genome sequences within 3 hours on a mid-class laptop. EBG is available under GNU GPL3. Data and Code Availabilitygithub.com/wiegertj/EBG github.com/wiegertj/EBG-train Contactjulius-wiegert@web.de

bioinformatics↗

bioRxiv2023

Simulations of sequence evolution: how (un)realistic they really are and why

MotivationSimulating Multiple Sequence Alignments (MSAs) using probabilistic models of sequence evolution plays an important role in the evaluation of phylogenetic inference tools, and is crucial to the development of novel learning-based approaches for phylogenetic reconstruction, for instance, neural networks. These models and the resulting simulated data need to be as realistic as possible to be indicative of the performance of the developed tools on empirical data and to ensure that neural networks trained on simulations perform well on empirical data. Over the years, numerous models of evolution have been published with the goal to represent as faithfully as possible the sequence evolution process and thus simulate empirical-like data. In this study, we simulated DNA and protein MSAs under increasingly complex models of evolution with and without insertion/deletion (indel) events using a state-of-the-art sequence simulator. We assessed their realism by quantifying how accurately supervised learning methods are able to predict whether a given MSA is simulated or empirical. ResultsOur results show that we can distinguish between empirical and simulated MSAs with high accuracy using two distinct and independently developed classification approaches across all tested models of sequence evolution. Our findings suggest that the current state-of-the-art models fail to accurately replicate several aspects of empirical MSAs, including site-wise rates as well as amino acid and nucleotide composition. Data and Code AvailabilityAll simulated and empirical MSAs, as well as all analysis results, are available at https://cme.h-its.org/exelixis/material/simulation_study.tar.gz. All scripts required to reproduce our results are available at https://github.com/tschuelia/SimulationStudy and https://github.com/JohannaTrost/seqsharp. Contactjulia.haag@h-its.org

bioinformatics↗

bioRxiv2023

Adaptive RAxML-NG: Accelerating Phylogenetic inference under Maximum Likelihood using dataset difficulty

MotivationPhylogenetic inferences under the Maximum-Likelihood (ML) criterion deploy heuristic tree search strategies to explore the vast search space. Depending on the input dataset, searches from different starting trees might all converge to a single tree topology. Often, though, distinct searches infer multiple topologies with large log-likelihood score differences or yield topologically highly distinct, yet almost equally likely, trees. Recently, Haag et al. introduced an approach to quantify, and implemented machine learning methods to predict, the difficulty of an MSA with respect to phylogenetic inference. Easy MSAs exhibit a single likelihood peak on their likelihood surface, associated with a single tree topology to which most, if not all, independent searches rapidly converge. However, as difficulty increases, multiple locally optimal likelihood peaks emerge, yet from highly distinct topologies. ResultsTo this end, we introduce and implement an adaptive tree search heuristic in RAxML-NG, which modifies the thoroughness of the tree search strategy as a function of the predicted difficulty. Our adaptive strategy is based upon three observations. First, on easy datasets, searches converge rapidly and can hence be terminated at an earlier stage. Second, over-analyzing difficult datasets is hopeless and, thus, it suffices to quickly infer only one of the numerous almost equally likely topologies, to reduce overall execution time. Third, more extensive searches are justified and required on datasets with intermediate difficulty. While the likelihood surface exhibits multiple locally optimal peaks in this case, a small proportion of them is significantly better. Our experimental results for the adaptive heuristic on 9, 515 empirical and 5, 000 simulated datasets with varying difficulty exhibit substantial speedups, especially on easy and difficult datasets (53% of total MSAs), where we observe average speedups of more than 10x. Further, approximately 94% of the inferred trees using the adaptive strategy are statistically indistinguishable from the trees inferred under the standard strategy (RAxML-NG). AvailabilityGNU GPL at https://github.com/togkousa/raxml-ng/tree/adaptive. ContactAnastasis.Togkousidis@h-its.org Supplementary MaterialAvailable

bioinformatics↗

bioRxiv2022

A representative Performance Assessment of Maximum Likelihood based Phylogenetic Inference Tools

SummaryThe evaluation of phylogenetic inference tools is commonly conducted on simulated and empirical sequence data alignments. An open question is how representative these alignments are with respect to those, commonly analyzed by users. Based upon the RAxMLGrove database, it is now possible to simulate DNA sequences based on more than 70, 000 representative RAxML and RAxML-NG tree inferences on empirical datasets conducted on the RAxML web servers. This allows to assess the phylogenetic tree inference accuracy of various inference tools based on realistic and representative simulated DNA alignments. We simulated 20, 000 MSAs based on representative datasets (in terms of signal strength) from RAxMLGrove, and used 5, 000 datasets from the TreeBASE database, to assess the inference accuracy of FastTree2, IQ-TREE2, and RAxML-NG. We find that on quantifiably difficult-to-analyze MSAs all of the analysed tools perform poorly, such that the quicker FastTree2, can constitute a viable alternative to infer trees. We also find, that there are substantial differences between accuracy results on simulated and empirical data, despite the fact that a substantial effort was undertaken to simulate sequences under as realistic as possible settings. ContactDimitri Hohler, dimitri.hoehler@h-its.org

bioinformatics↗

bioRxiv2022

The Free Lunch is not over yet - Systematic Exploration of Numerical Thresholds in Phylogenetic Inference

Maximum Likelihood (ML) is a widely used phylogenetic inference model. ML implementations heavily rely on numerical optimization routines that use internal numerical thresholds to determine convergence. We systematically analyze the impact of these threshold settings on the log-likelihood and runtimes for ML tree inferences with RAxML-NG, IQ-TREE, and FastTree on empirical datasets. We provide empirical evidence that we can substantially accelerate tree inferences with RAxML-NG and IQ-TREE by changing the default values of two such numerical thresholds. At the same time, altering these settings does not significantly impact the quality of the inferred trees. We further show that increasing both thresholds accelerates the RAxML-NG bootstrap without influencing the resulting support values. For RAxML-NG, increasing the likelihood thresholds{epsilon} LnL and{epsilon} brlen to 10 and 103 respectively results in an average tree inference speedup of 1.9 {+/-} 0.6 on Data collection 1, 1.8 {+/-} 1.1 on Data collection 2, and 1.9 {+/-} 0.8 on Data collection 2 for the RAxML-NG bootstrap. Increasing the likelihood threshold{epsilon} LnL to 10 in IQ-TREE results in an average tree inference speedup of 1.3 {+/-} 0.4 on Data collection 1 and 1.3 {+/-} 0.9 on Data collection 2.

bioinformatics↗

bioRxiv2022

Voltage to Calcium Transformation Enhances Direction Selectivity in Drosophila T4 neurons

A critical step in neural information processing is the transformation of membrane voltage into calcium signals leading to transmitter release. However, the effect of voltage to calcium transformation on neural responses to different sensory stimuli is not well understood. Here, we use in vivo two-photon imaging of genetically encoded voltage and calcium indicators, Arclight and GCaMP6f respectively, to measure responses in Drosophila direction-selective T4 neurons. Comparison between Arclight and GCaMP6f signals revealed calcium signals to have a significantly higher direction selectivity compared to voltage signals. Using these recordings we build a model which transforms T4 voltage responses to calcium responses. The model reproduces experimentally measured calcium responses across different visual stimuli using different temporal filtering steps and a stationary non-linearity. These findings provide a mechanistic underpinning of the voltage-to-calcium transformation and show how this processing step, in addition to synaptic mechanisms on the dendrites of T4 cells, enhances direction selectivity in the output signal of T4 neurons.

neuroscience↗

bioRxiv2022

From Easy to Hopeless - Predicting the Difficulty of Phylogenetic Analyses

Phylogenetic analyses under the Maximum Likelihood model are time and resource intensive. To adequately capture the vastness of tree space, one needs to infer multiple independent trees. On some datasets, multiple tree inferences converge to similar tree topologies, on others to multiple, topologically highly distinct yet statistically indistinguishable topologies. At present, no method exists to quantify and predict this behavior. We introduce a method to quantify the degree of difficulty for analyzing a dataset and present Pythia, a Random Forest Regressor that accurately predicts this difficulty. Pythia predicts the degree of difficulty of analyzing a dataset prior to initiating Maximum Likelihood based tree inferences. Pythia can be used to increase user awareness with respect to the amount of signal and uncertainty to be expected in phylogenetic analyses, and hence inform an appropriate (post-)analysis setup. Further, it can be used to select appropriate search algorithms for easy-, intermediate-, and hard-to-analyze datasets.

bioinformatics↗

bioRxiv2017

siRNA biogenesis, DNA methylation and target locus silencing are distinct and separable Pol IV sub-functions differentially dependent on the largest subunit's CTD

Plant nuclear multisubunit RNA polymerase IV plays a key role in the RNA-directed DNA methylation (RdDM) pathway for transcriptional silencing of transposons, viruses and specific genes by synthesizing precursors of 24 nt siRNAs that guide the process. The Pol IV largest subunit, NRPD1 is derived from the Pol II largest subunit but has a unique carboxy-terminal domain (CTD) of unknown function. We show that the NRPD1 CTD is critical for transcriptional silencing of target loci and for producing 24 nt siRNAs at high levels. However, the CTD is surprisingly dispensable for near wild-type levels of Pol IV-dependent genomic cytosine methylation. These results suggest that low levels of 24 nt siRNAs, produced at only 20-30% of wild-type levels, are sufficient for full RNA-directed DNA methylation, yet insufficient for silencing, suggesting additional roles for siRNAs beyond DNA methylation. Moreover, at a subset of target loci, neither siRNA levels nor cytosine methylation are impaired upon deletion of the CTD, yet silencing is lost. Collectively, the non-linear relationships between siRNA levels, cytosine methylation and silencing suggest the existence of additional mechanisms of silencing dependent on Pol IV transcription and mediated by the CTD, such as promoter occlusion to inhibit the activities of other polymerases.

molecular biology↗

bioRxiv2017

Subdomains of the Pol V largest subunit CTD mediate locus-specific RNA-directed DNA methylation

Plant multisubunit RNA Polymerase V transcription recruits Argonaute siRNA complexes that specify sites of RNA-directed DNA methylation (RdDM) for gene silencing. Pol V's largest subunit, NRPE1, evolved from the largest subunit of Pol II but has a distinctive carboxyl-terminal domain (CTD). We show that the Pol V CTD is dispensable for catalytic activity in vitro, yet essential in vivo. One CTD subdomain (DeCL), is required for Pol V function at virtually all loci. other CTD subdomains have locusspecific effects. In a yeast two-hybrid screen, the 3'->5' exoribonuclease, RRP6L1 was identified as an interactor with the DeCL subdomain and DeCL and glutamine-serine-rich (QS) subdomains, located downstream from an Argonaute-binding repeat subdomain. Experimental evidence indicates that RRP6L1 trims the 3 ends of Pol V transcripts sliced by ARGONAUTE 4 (AGO4), suggesting a model whereby the CTD enables the spatial and temporal coordination of AGO4 and RRP6L1 RNA processing activities.

molecular biology↗