Lerna: transformer architectures for configuring error correction tools for short- and long-read genome sequencing

Background Sequencing technologies are prone to errors, making error correction (EC) necessary for downstream applications. EC tools need to be manually configured for optimal performance. We find that the optimal parameters (e.g., k-mer size) are both tool- and dataset-dependent. Moreover, evaluating the performance (i.e., Alignment-rate or Gain) of a given tool usually relies on a reference genome, but quality reference genomes are not always available. We introduce Lerna for the automated configuration of k-mer-based EC tools. Lerna first creates a language model (LM) of the uncorrected genomic reads, and then, based on this LM, calculates a metric called the perplexity metric to evaluate the corrected reads for different parameter choices. Next, it finds the one that produces the highest alignment rate without using a reference genome. The fundamental intuition of our approach is that the perplexity metric is inversely correlated with the quality of the assembly after error correction. Therefore, Lerna leverages the perplexity metric for automated tuning of k-mer sizes without needing a reference genome. Results First, we show that the best k-mer value can vary for different datasets, even for the same EC tool. This motivates our design that automates k-mer size selection without using a reference genome. Second, we show the gains of our LM using its component attention-based transformers. We show the model’s estimation of the perplexity metric before and after error correction. The lower the perplexity after correction, the better the k-mer size. We also show that the alignment rate and assembly quality computed for the corrected reads are strongly negatively correlated with the perplexity, enabling the automated selection of k-mer values for better error correction, and hence, improved assembly quality. We validate our approach on both short and long reads. Additionally, we show that our attention-based models have significant runtime improvement for the entire pipeline—18$$\times$$ × faster than previous works, due to parallelizing the attention mechanism and the use of JIT compilation for GPU inferencing. Conclusion Lerna improves de novo genome assembly by optimizing EC tools. Our code is made available in a public repository at: https://github.com/icanforce/lerna-genomics.

Rumination Derails Reinforcement Learning With Possible Implications for Ineffective Behavior

How does rumination affect reinforcement learning—the ubiquitous process by which people adjust behavior after error to behave more effectively in the future? In a within-subjects design ( N = 49), we tested whether experimentally manipulated rumination disrupts reinforcement learning in a multidimensional learning task previously shown to rely on selective attention. Rumination impaired performance, yet unexpectedly, this impairment could not be attributed to decreased attentional breadth (quantified using a decay parameter in a computational model). Instead, trait rumination (between subjects) was associated with higher decay rates (implying narrower attention) but not with impaired performance. Our task-performance results accord with the possibility that state rumination promotes stress-generating behavior in part by disrupting reinforcement learning. The trait-rumination finding accords with the predictions of a prominent model of trait rumination (the attentional-scope model). More work is needed to understand the specific mechanisms by which state rumination disrupts reinforcement learning.

Multicenter clinical evaluation of VITEK® 2 meropenem-vaborbactam for susceptibility testing of Enterobacterales and Pseudomonas aeruginosa

The carbapenem/beta-lactamase inhibitor (meropenem-vaborbactam; MEV) used to treat complicated urinary tract infections and pyelonephritis in adults was approved in 2017 by the U.S. Food and Drug Administration (FDA). We evaluated VITEK 2 MEV (bioMérieux, Durham, NC) compared to the reference broth microdilution (BMD) method. Of 449 Enterobacterales isolates analyzed per FDA/CLSI breakpoints, overall performance was 98.2% Essential Agreement (EA), 98.7% Category Agreement (CA), and 0% Very Major Errors (VME) or Major Errors (ME). For FDA intended for use 438 Enterobacterales isolates, performance was 98.2% EA, 98.6% CA, and 0% VME or ME. Evaluable EA was 81.0% but with only 42 on-scale evaluable results. Individual species demonstrated EA and CA rates ≥ 90% without any VME or ME. When evaluated using European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoints, overall VITEK 2 MEV performance for Enterobacterales and Pseudomonas aeruginosa demonstrated 97.3% EA, 99.2% CA, 2.3% VME, and 0.6% ME (after error resolution: 97.3% EA, 99.4% CA, 2.2% VME, and 0.4% ME) compared to the reference BMD method. Performance for P. aeruginosa included 92.2% EA, 97.4% CA, 0% VME, and 3.0% ME (after error resolution: 92.2% EA, 98.7% CA, 0% VME, and 1.5% ME). Performance for Enterobacterales included 98.2% EA, 99.6% CA, 3.0% VME, and 0.2% ME. Evaluable EA was 80.6% but due to only 67 evaluable results. These findings support VITEK 2 MEV as an accurate automated system for MEV susceptibility testing of Enterobacterales and P. aeruginosa and could be an alternate solution to the manual labor intensive reference BMD method.

Inter-player variability in game patterns in high-level women's volleyball: A study with Outside Hitters (Near vs. Away) using Social Network Analysis. [Variabilidad entre jugadores en los patrones de juego en el voleibol femenino de alto nivel: Un estudio con bateadores externos (Cerca versus Lejos) usando Análisis de Redes Sociales].

A wide body of research on team sports has focused on positional status based differences, providing information on inter-player variability according to the functional roles within the game. However, research addressing inter-player variability within the same positional/function status is scarce. The present article presents an analysis of inter-player variability within the same positional status during critical moments, in high-level women's volleyball, using Social Network Analysis. Attack actions of the outside hitters near (OHN) and away (OHA) from the setter were analysed in ten matches from the 2019 Volleyball Nations League Finals (268 plays). Two independent Eigenvector Centrality networks were created, one for OHN and another for OHA. Main results: (a) in side-out with ideal setting conditions, the OHA used more tips and exploration of the block than the OHN; under non-ideal setting conditions, the OHN had slower attack tempos than the OHA; (b) OHA used tip and directed attacks after error situations while OHN was typically not requested after error situations; (c) in transition, OHN typically attacked after having performed a previous action, performing a dual task within each ball possession, while OHA only attacked when there was no prior action; (d) there were also inter-positional similarities, with both OHN and OHA preferring a strong attack in ideal conditions during KI and KIV, and slower tempos in transition in non-ideal conditions. Conclusions: Even within the same positional status, there seems to be subtle, but relevant inter-player variability. Consequently, coaches should devote careful attention when assigning players to positional.

Pre-training RNNs on ecologically relevant tasks explains sub-optimal behavioral reset

When faced with a new task, animals′ cognitive capabilities are determined both by individual experience and by structural priors evolved to leverage the statistics of natural environments. Rats can quickly learn to capitalize on the trial sequence correlations of two-alternative forced choice (2AFC) tasks after correct trials, but consistently deviate from optimal behavior after error trials, when they waive the accumulated evidence. To understand this outcome-dependent gating, we first show that Recurrent Neural Networks (RNNs) trained in the same 2AFC task outperform animals as they can readily learn to use previous trials′ information both after correct and error trials. We hypothesize that, while RNNs can optimize their behavior in the 2AFC task without a priori restrictions, rats′ strategy is constrained by a structural prior adapted to a natural environment in which rewarded and non-rewarded actions provide largely asymmetric information. When pre-training RNNs in a more ecological task with more than two possible choices, networks develop a strategy by which they gate off the across-trial evidence after errors, mimicking rats′ behavior. Our results suggest that the observed suboptimal behavior reflects the influence of a structural prior that, adaptive in a natural multi-choice environment, constrains performance in a 2AFC laboratory task.

Rumination Derails Reinforcement Learning with Possible Implications for Ineffective Behavior

How does rumination affect reinforcement learning—the ubiquitous process by which we adjust behavior after error in order to behave more effectively in the future? In a within-subject design (n=49), we tested whether experimentally induced rumination disrupts reinforcement learning in a multidimensional learning task previously shown to rely on selective attention. Rumination impaired performance, yet unexpectedly this impairment could not be attributed to decreased attentional breadth (quantified using a “decay” parameter in a computational model). Instead, trait rumination (between subjects) was associated with higher decay rates (implying narrower attention), yet not with impaired performance. Our task-performance results accord with the possibility that state rumination promotes stress-generating behavior in part by disrupting reinforcement learning. The trait-rumination finding accords with the predictions of a prominent model of trait rumination (the attentional-scope model). More work is needed to understand the specific mechanisms by which state rumination disrupts reinforcement learning.

Once Again on the Subject of the Great Caravel "Peter von Danzig"

As can be seen from the comments herein, every time that Długokęcki tries to add something new to the main themes I deal with in writing the history of the caravel, he makes error after error. It applies to both the marine layer of monograph and understanding of the European context. His interpretation of the sources and the theories he builds on this basis in order to create an alternative picture are unsuccessful. All in all, though it is evident that he has tried very hard, Długokęcki is unable to change any of the findings regarding the major themes addressed in my monograph.

Comparative Study of Heart Rate, Blood Pressure and Selective Attention of Subjects Before and After Music

Introduction: Studies have shown different views regarding the effect of music in vitals e.g Heart rate (HR), Blood pressure (BP) and atiention. The effect of preferred music with lyrics in vitals and reaction time in stroop test has not been performed in Nepalese students so, we conducted the study. Objective: To find out the change in HR, BP and reaction time in Stroop test before and after their preferred music with lyrics. Methodology Thirty male medical and paramedical students aged 25.27 ± 2.0 participated in study. The vital signs and reaction time in Stroop test before and after music was taken. Results Paired-t test was used to compare means before and after exposure to music. The means are expressed as Mean ± SD. Heart rate (HR) increased after exposure to music (66.33±9.51 Vs 67.2±8.44) (p<.05). The error in Stroop test was less after music (.66±.49 Vs.63±.66) (p<.05). The reaction time after error correction decreased post exposure to music (24.117±4.61Vs23.29±4.45) (p<.05). Conclusion The heart rate increased after exposure to music. The errors decreased after listening to music which also decreased reaction time.