High dimensional predictions of suicide risk in 4.2 million US Veterans using ensemble transfer learning.

We present an ensemble transfer learning method to predict suicide from Veterans Affairs (VA) electronic medical records (EMR). A diverse set of base models was trained to predict a binary outcome constructed from reported suicide, suicide attempt, and overdose diagnoses with varying choices of study design and prediction methodology. Each model used twenty cross-sectional and 190 longitudinal variables observed in eight time intervals covering 7.5 years prior to the time of prediction. Ensembles of seven base models were created and fine-tuned with ten variables expected to change with study design and outcome definition in order to predict suicide and combined outcome in a prospective cohort. The ensemble models achieved c-statistics of 0.73 on 2-year suicide risk and 0.83 on the combined outcome when predicting on a prospective cohort of [Formula: see text] 4.2 M veterans. The ensembles rely on nonlinear base models trained using a matched retrospective nested case-control (Rcc) study cohort and show good calibration across a diversity of subgroups, including risk strata, age, sex, race, and level of healthcare utilization. In addition, a linear Rcc base model provided a rich set of biological predictors, including indicators of suicide, substance use disorder, mental health diagnoses and treatments, hypoxia and vascular damage, and demographics.

Prevention efficacy of the broadly neutralizing antibody VRC01 depends on HIV-1 envelope sequence features.

In the Antibody Mediated Prevention (AMP) trials (HVTN 704/HPTN 085 and HVTN 703/HPTN 081), prevention efficacy (PE) of the monoclonal broadly neutralizing antibody (bnAb) VRC01 (vs. placebo) against HIV-1 acquisition diagnosis varied according to the HIV-1 Envelope (Env) neutralization sensitivity to VRC01, as measured by 80% inhibitory concentration (IC80). Here, we performed a genotypic sieve analysis, a complementary approach to gaining insight into correlates of protection that assesses how PE varies with HIV-1 sequence features. We analyzed HIV-1 Env amino acid (AA) sequences from the earliest available HIV-1 RNA-positive plasma samples from AMP participants diagnosed with HIV-1 and identified Env sequence features that associated with PE. The strongest Env AA sequence correlate in both trials was VRC01 epitope distance that quantifies the divergence of the VRC01 epitope in an acquired HIV-1 isolate from the VRC01 epitope of reference HIV-1 strains that were most sensitive to VRC01-mediated neutralization. In HVTN 704/HPTN 085, the Env sequence-based predicted probability that VRC01 IC80 against the acquired isolate exceeded 1 µg/mL also significantly associated with PE. In HVTN 703/HPTN 081, a physicochemical-weighted Hamming distance across 50 VRC01 binding-associated Env AA positions of the acquired isolate from the most VRC01-sensitive HIV-1 strain significantly associated with PE. These results suggest that incorporating mutation scoring by BLOSUM62 and weighting by the strength of interactions at AA positions in the epitope:VRC01 interface can optimize performance of an Env sequence-based biomarker of VRC01 prevention efficacy. Future work could determine whether these results extend to other bnAbs and bnAb combinations.

Deep learning uncertainty quantification for clinical text classification.

INTRODUCTION: Machine learning algorithms are expected to work side-by-side with humans in decision-making pipelines. Thus, the ability of classifiers to make reliable decisions is of paramount importance. Deep neural networks (DNNs) represent the state-of-the-art models to address real-world classification. Although the strength of activation in DNNs is often correlated with the network's confidence, in-depth analyses are needed to establish whether they are well calibrated. METHOD: In this paper, we demonstrate the use of DNN-based classification tools to benefit cancer registries by automating information extraction of disease at diagnosis and at surgery from electronic text pathology reports from the US National Cancer Institute (NCI) Surveillance, Epidemiology, and End Results (SEER) population-based cancer registries. In particular, we introduce multiple methods for selective classification to achieve a target level of accuracy on multiple classification tasks while minimizing the rejection amount-that is, the number of electronic pathology reports for which the model's predictions are unreliable. We evaluate the proposed methods by comparing our approach with the current in-house deep learning-based abstaining classifier. RESULTS: Overall, all the proposed selective classification methods effectively allow for achieving the targeted level of accuracy or higher in a trade-off analysis aimed to minimize the rejection rate. On in-distribution validation and holdout test data, with all the proposed methods, we achieve on all tasks the required target level of accuracy with a lower rejection rate than the deep abstaining classifier (DAC). Interpreting the results for the out-of-distribution test data is more complex; nevertheless, in this case as well, the rejection rate from the best among the proposed methods achieving 97% accuracy or higher is lower than the rejection rate based on the DAC. CONCLUSIONS: We show that although both approaches can flag those samples that should be manually reviewed and labeled by human annotators, the newly proposed methods retain a larger fraction and do so without retraining-thus offering a reduced computational cost compared with the in-house deep learning-based abstaining classifier.

Pion-Nucleon Sigma Term from Lattice QCD.

We present an analysis of the pion-nucleon σ-term σ_{πN} using six ensembles with 2+1+1-flavor highly improved staggered quark action generated by the MILC Collaboration. The most serious systematic effect in lattice calculations of nucleon correlation functions is the contribution of excited states. We estimate these using chiral perturbation theory (χPT) and show that the leading contribution to the isoscalar scalar charge comes from Nπ and Nππ states. Therefore, we carry out two analyses of lattice data to remove excited-state contamination, the standard one and a new one including Nπ and Nππ states. We find that the standard analysis gives σ_{πN}=41.9(4.9) MeV, consistent with previous lattice calculations, while our preferred χPT-motivated analysis gives σ_{πN}=59.6(7.4) MeV, which is consistent with phenomenological values obtained using πN scattering data. Our data on one physical pion mass ensemble were crucial for exposing this difference, therefore, calculations on additional physical mass ensembles are needed to confirm our result and resolve the tension between lattice QCD and phenomenology.

HIV-1 and SARS-CoV-2: Patterns in the evolution of two pandemic pathogens.

Humanity is currently facing the challenge of two devastating pandemics caused by two very different RNA viruses: HIV-1, which has been with us for decades, and SARS-CoV-2, which has swept the world in the course of a single year. The same evolutionary strategies that drive HIV-1 evolution are at play in SARS-CoV-2. Single nucleotide mutations, multi-base insertions and deletions, recombination, and variation in surface glycans all generate the variability that, guided by natural selection, enables both HIV-1's extraordinary diversity and SARS-CoV-2's slower pace of mutation accumulation. Even though SARS-CoV-2 diversity is more limited, recently emergent SARS-CoV-2 variants carry Spike mutations that have important phenotypic consequences in terms of both antibody resistance and enhanced infectivity. We review and compare how these mutational patterns manifest in these two distinct viruses to provide the variability that fuels their evolution by natural selection.

Qubit Regularization of Asymptotic Freedom.

We provide strong evidence that the asymptotically free (1+1)-dimensional nonlinear O(3) sigma model can be regularized using a quantum lattice Hamiltonian, referred to as the "Heisenberg comb," that acts on a Hilbert space with only two qubits per spatial lattice site. The Heisenberg comb consists of a spin-half antiferromagnetic Heisenberg chain coupled antiferromagnetically to a second local spin-half particle at every lattice site. Using a world-line Monte Carlo method, we show that the model reproduces the universal step-scaling function of the traditional model up to correlation lengths of 200000 in lattice units and argue how the continuum limit could emerge. We provide a quantum circuit description of the time evolution of the model and argue that near-term quantum computers may suffice to demonstrate asymptotic freedom.

Tracking Changes in SARS-CoV-2 Spike: Evidence that D614G Increases Infectivity of the COVID-19 Virus.

A SARS-CoV-2 variant carrying the Spike protein amino acid change D614G has become the most prevalent form in the global pandemic. Dynamic tracking of variant frequencies revealed a recurrent pattern of G614 increase at multiple geographic levels: national, regional, and municipal. The shift occurred even in local epidemics where the original D614 form was well established prior to introduction of the G614 variant. The consistency of this pattern was highly statistically significant, suggesting that the G614 variant may have a fitness advantage. We found that the G614 variant grows to a higher titer as pseudotyped virions. In infected individuals, G614 is associated with lower RT-PCR cycle thresholds, suggestive of higher upper respiratory tract viral loads, but not with increased disease severity. These findings illuminate changes important for a mechanistic understanding of the virus and support continuing surveillance of Spike mutations to aid with development of immunological interventions.

Axial Vector Form Factors from Lattice QCD that Satisfy the PCAC Relation.

Previous lattice QCD calculations of axial vector and pseudoscalar form factors show significant deviation from the partially conserved axial current (PCAC) relation between them. Since the original correlation functions satisfy PCAC, the observed deviations from the operator identity cast doubt on whether all of the systematics in the extraction of form factors from the correlation functions are under control. We identify the problematic systematic as a missed excited state, whose energy as a function of the momentum transfer squared Q^{2} is determined from the analysis of the three-point functions themselves. Its energy is much smaller than those of the excited states previously considered, and including it impacts the extraction of all of the ground state matrix elements. The form factors extracted using these mass and energy gaps satisfy PCAC and another consistency condition, and they validate the pion-pole dominance hypothesis. We also show that the extraction of the axial charge g_{A} is very sensitive to the value of the mass gaps of the excited states used, and current lattice data do not provide an unambiguous determination of these, unlike the Q^{2}≠0 case. To highlight the differences and improvement between the conventional vs the new analysis strategy, we present a comparison of results obtained on a physical pion mass ensemble at a≈0.0871 fm. With the new strategy, we find g_{A}=1.30(6) and axial charge radius r_{A}=0.74(6) fm, both extracted using the z expansion to parametrize the Q^{2} behavior of G_{A}(Q^{2}), and g_{P}^{*}=8.06(44), obtained using the pion-pole dominance ansatz to fit the Q^{2} behavior of the induced pseudoscalar form factor G[over ˜]_{P}(Q^{2}). These results are consistent with current phenomenological values.

Estimating the Timing of Early Simian-Human Immunodeficiency Virus Infections: a Comparison between Poisson Fitter and BEAST.

Many HIV prevention strategies are currently under consideration where it is highly informative to know the study participants' times of infection. These can be estimated using viral sequence data sampled early in infection. However, there are several scenarios that, if not addressed, can skew timing estimates. These include multiple transmitted/founder (TF) viruses, APOBEC (apolipoprotein B mRNA editing enzyme, catalytic polypeptide-like)-mediated mutational enrichment, and recombination. Here, we suggest a pipeline to identify these problems and resolve the biases that they introduce. We then compare two modeling strategies to obtain timing estimates from sequence data. The first, Poisson Fitter (PF), is based on a Poisson model of random accumulation of mutations relative to the TF virus (or viruses) that established the infection. The second uses a coalescence-based phylogenetic strategy as implemented in BEAST. The comparison is based on timing predictions using plasma viral RNA (cDNA) sequence data from 28 simian-human immunodeficiency virus (SHIV)-infected animals for which the exact day of infection is known. In this particular setting, based on nucleotide sequences from samples obtained in early infection, the Poisson method yielded more accurate, more precise, and unbiased estimates for the time of infection than did the explored implementations of BEAST.IMPORTANCE The inference of the time of infection is a critical parameter in testing the efficacy of clinical interventions in protecting against HIV-1 infection. For example, in clinical trials evaluating the efficacy of passively delivered antibodies (Abs) for preventing infections, accurate time of infection data are essential for discerning levels of the Abs required to confer protection, given the natural Ab decay rate in the human body. In such trials, genetic sequences from early in the infection are regularly sampled from study participants, generally prior to immune selection, when the viral population is still expanding and genetic diversity is low. In this particular setting of early viral growth, the Poisson method is superior to the alternative approach based on coalescent methods. This approach can also be applied in human vaccine trials, where accurate estimates of infection times help ascertain if vaccine-elicited immune protection wanes over time.

Accumulated mutations by 6 months of infection collectively render transmitted/founder HIV-1 significantly less fit.

OBJECTIVE: Viral fitness plays an important role in HIV-1 evolution, transmission and pathogenesis. However, how mutations accumulated during early infection affect viral fitness has not been well studied. METHODS: Paired infectious molecular clones (IMCs) for transmitted/founder (T/F) and 6-month (6-mo) viruses post infection were generated from 10 infected individuals to investigate the impact of accumulated mutations on viral fitness by comparing 6-mo viruses to their cognate T/F viruses. RESULTS: All ten 6-mo viruses were less fit than their cognate T/F viruses. Moreover, the fitness losses of the 6-mo viruses correlated with the decrease in viral loads from the peak of viremia. CONCLUSION: These results show that the mutations accumulated during half a year post infection collectively reduce viral fitness and thereby contribute to lowering viral loads.

AI Meets Exascale Computing: Advancing Cancer Research With Large-Scale High Performance Computing.

The application of data science in cancer research has been boosted by major advances in three primary areas: (1) Data: diversity, amount, and availability of biomedical data; (2) Advances in Artificial Intelligence (AI) and Machine Learning (ML) algorithms that enable learning from complex, large-scale data; and (3) Advances in computer architectures allowing unprecedented acceleration of simulation and machine learning algorithms. These advances help build in silico ML models that can provide transformative insights from data including: molecular dynamics simulations, next-generation sequencing, omics, imaging, and unstructured clinical text documents. Unique challenges persist, however, in building ML models related to cancer, including: (1) access, sharing, labeling, and integration of multimodal and multi-institutional data across different cancer types; (2) developing AI models for cancer research capable of scaling on next generation high performance computers; and (3) assessing robustness and reliability in the AI models. In this paper, we review the National Cancer Institute (NCI) -Department of Energy (DOE) collaboration, Joint Design of Advanced Computing Solutions for Cancer (JDACS4C), a multi-institution collaborative effort focused on advancing computing and data technologies to accelerate cancer research on three levels: molecular, cellular, and population. This collaboration integrates various types of generated data, pre-exascale compute resources, and advances in ML models to increase understanding of basic cancer biology, identify promising new treatment options, predict outcomes, and eventually prescribe specialized treatments for patients with cancer.

Intragenomic conflict over bet-hedging.

Genomic imprinting, where an allele's expression pattern depends on its parental origin, is thought to result primarily from an intragenomic evolutionary conflict. Imprinted genes are widely expressed in the brain and have been linked to various phenotypes, including behaviours related to risk tolerance. In this paper, we analyse a model of evolutionary bet-hedging in a system with imprinted gene expression. Previous analyses of bet-hedging have shown that natural selection may favour alleles and traits that reduce reproductive variance, even at the expense of reducing mean reproductive success, with the trade-off between mean and variance depending on the population size. In species where the sexes have different reproductive variances, this bet-hedging trade-off differs between maternally and paternally inherited alleles. Where males have the higher reproductive variance, alleles are more strongly selected to reduce variance when paternally inherited than when maternally inherited. We connect this result to phenotypes connected with specific imprinted genes, including delay discounting and social dominance. The empirical patterns are consistent with paternally expressed imprinted genes promoting risk-averse behaviours that reduce reproductive variance. Conversely, maternally expressed imprinted genes promote risk-tolerant, variance-increasing behaviours. We indicate how future research might further test the hypotheses suggested by our analysis. This article is part of the theme issue 'Risk taking and impulsive behaviour: fundamental discoveries, theoretical perspectives and clinical implications'.

CAT: computer aided triage improving upon the Bayes risk through ε-refusal triage rules.

BACKGROUND: Manual extraction of information from electronic pathology (epath) reports to populate the Surveillance, Epidemiology, and End Result (SEER) database is labor intensive. Systematizing the data extraction automatically using machine-learning (ML) and natural language processing (NLP) is desirable to reduce the human labor required to populate the SEER database and to improve the timeliness of the data. This enables scaling up registry efficiency and collection of new data elements. To ensure the integrity, quality, and continuity of the SEER data, the misclassification error of ML and NPL algorithms needs to be negligible. Current algorithms fail to achieve the precision of human experts who can bring additional information in their assessments. Differences in registry format and the desire to develop a common information extraction platform further complicate the ML/NLP tasks. The purpose of our study is to develop triage rules to partially automate registry workflow to improve the precision of the auto-extracted information. RESULTS: This paper presents a mathematical framework to improve the precision of a classifier beyond that of the Bayes classifier by selectively classifying item that are most likely to be correct. This results in a triage rule that only classifies a subset of the item. We characterize the optimal triage rule and demonstrate its usefulness in the problem of classifying cancer site from electronic pathology reports to achieve a desired precision. CONCLUSIONS: From the mathematical formalism, we propose a heuristic estimate for triage rule based on post-processing the soft-max output from standard machine learning algorithms. We show, in test cases, that the triage rule significantly improve the classification accuracy.

Superinfection and cure of infected cells as mechanisms for hepatitis C virus adaptation and persistence.

RNA viruses exist as a genetically diverse quasispecies with extraordinary ability to adapt to abrupt changes in the host environment. However, the molecular mechanisms that contribute to their rapid adaptation and persistence in vivo are not well studied. Here, we probe hepatitis C virus (HCV) persistence by analyzing clinical samples taken from subjects who were treated with a second-generation HCV protease inhibitor. Frequent longitudinal viral load determinations and large-scale single-genome sequence analyses revealed rapid antiviral resistance development, and surprisingly, dynamic turnover of dominant drug-resistant mutant populations long after treatment cessation. We fitted mathematical models to both the viral load and the viral sequencing data, and the results provided strong support for the critical roles that superinfection and cure of infected cells play in facilitating the rapid turnover and persistence of viral populations. More broadly, our results highlight the importance of considering viral dynamics and competition at the intracellular level in understanding rapid viral adaptation. Thus, we propose a theoretical framework integrating viral and molecular mechanisms to explain rapid viral evolution, resistance, and persistence despite antiviral treatment and host immune responses.

Tracking HIV-1 recombination to resolve its contribution to HIV-1 evolution in natural infection.

Recombination in HIV-1 is well documented, but its importance in the low-diversity setting of within-host diversification is less understood. Here we develop a novel computational tool (RAPR (Recombination Analysis PRogram)) to enable a detailed view of in vivo viral recombination during early infection, and we apply it to near-full-length HIV-1 genome sequences from longitudinal samples. Recombinant genomes rapidly replace transmitted/founder (T/F) lineages, with a median half-time of 27 days, increasing the genetic complexity of the viral population. We identify recombination hot and cold spots that differ from those observed in inter-subtype recombinants. Furthermore, RAPR analysis of longitudinal samples from an individual with well-characterized neutralizing antibody responses shows that recombination helps carry forward resistance-conferring mutations in the diversifying quasispecies. These findings provide insight into molecular mechanisms by which viral recombination contributes to HIV-1 persistence and immunopathogenesis and have implications for studies of HIV transmission and evolution in vivo.

Kinship structures create persistent channels for language transmission.

Languages are transmitted through channels created by kinship systems. Given sufficient time, these kinship channels can change the genetic and linguistic structure of populations. In traditional societies of eastern Indonesia, finely resolved cophylogenies of languages and genes reveal persistent movements between stable speech communities facilitated by kinship rules. When multiple languages are present in a region and postmarital residence rules encourage sustained directional movement between speech communities, then languages should be channeled along uniparental lines. We find strong evidence for this pattern in 982 individuals from 25 villages on two adjacent islands, where different kinship rules have been followed. Core groups of close relatives have stayed together for generations, while remaining in contact with, and marrying into, surrounding groups. Over time, these kinship systems shaped their gene and language phylogenies: Consistently following a postmarital residence rule turned social communities into speech communities.

A strongly selected mutation in the HIV-1 genome is independent of T cell responses and neutralizing antibodies.

BACKGROUND: Mutations rapidly accumulate in the HIV-1 genome after infection. Some of those mutations are selected by host immune responses and often cause viral fitness losses. This study is to investigate whether strongly selected mutations that are not associated with immune responses result in fitness losses. RESULTS: Strongly selected mutations were identified by analyzing 5'-half HIV-1 genome (gag/pol) sequences from longitudinal samples of subject CH0131. The K43R mutation in the gag gene was first detected at day 91 post screening and was fixed in the viral population at day 273 while the synonymous N323tc mutation was first detected at day 177 and fixed at day 670. No conventional or cryptic T cell responses were detected against either mutation sites by ELISpot analysis. However, when fitness costs of both mutations were measured by introducing each mutation into their cognate transmitted/founder (T/F) viral genome, the K43R mutation caused a significant fitness loss while the N323tc mutation had little impact on viral fitness. CONCLUSIONS: The rapid fixation, the lack of detectable immune responses and the significant fitness cost of the K43R mutation suggests that it was strongly selected by host factors other than T cell responses and neutralizing antibodies.

The Emergence of Life as a First-Order Phase Transition.

It is well known that life on Earth alters its environment over evolutionary and geological timescales. An important open question is whether this is a result of evolutionary optimization or a universal feature of life. In the latter case, the origin of life would be coincident with a shift in environmental conditions. Here we present a model for the emergence of life in which replicators are explicitly coupled to their environment through the recycling of a finite supply of resources. The model exhibits a dynamic, first-order phase transition from nonlife to life, where the life phase is distinguished by selection on replicators. We show that environmental coupling plays an important role in the dynamics of the transition. The transition corresponds to a redistribution of matter in replicators and their environment, driven by selection on replicators, exhibiting an explosive growth in diversity as replicators are selected. The transition is accurately tracked by the mutual information shared between replicators and their environment. In the absence of successfully repartitioning system resources, the transition fails to complete, leading to the possibility of many frustrated trials before life first emerges. Often, the replicators that initiate the transition are not those that are ultimately selected. The results are consistent with the view that life's propensity to shape its environment is indeed a universal feature of replicators, characteristic of the transition from nonlife to life. We discuss the implications of these results for understanding life's emergence and evolutionary transitions more broadly. Key Words: Origin of life-Prebiotic evolution-Astrobiology-Biopolymers-Life. Astrobiology 17, 266-276.

Transmission of Multiple HIV-1 Subtype C Transmitted/founder Viruses into the Same Recipients Was not Determined by Modest Phenotypic Differences.

A severe bottleneck exists during HIV-1 mucosal transmission. However, viral properties that determine HIV-1 transmissibility are not fully elucidated. We identified multiple transmitted/founder (T/F) viruses in six HIV-1-infected subjects by analyzing whole genome sequences. Comparison of biological phenotypes of different T/F viruses from the same individual allowed us to more precisely identify critical determinants for viral transmissibility since they were transmitted under similar conditions. All T/F viruses used coreceptor CCR5, while no T/F viruses used CXCR4 or GPR15. However, the efficiency for different T/F viruses from the same individual to use CCR5 was significantly variable, and the differences were even more significant for usage of coreceptors FPRL1, CCR3 and APJ. Resistance to IFN-α was also different between T/F viruses in 2 of 3 individuals. The relative fitness between T/F viruses from the same subject was highly variable (2-6%). Importantly, the levels of coreceptor usage efficiency, resistance to IFN-α and viral fitness were not associated with proportions of T/F viruses in each individual during acute infection. Our results show that the modest but significant differences in coreceptor usage efficiency, IFN-α sensitivity and viral fitness each alone may not play a critical role in HIV-1 transmission.

Optimal Combinations of Broadly Neutralizing Antibodies for Prevention and Treatment of HIV-1 Clade C Infection.

The identification of a new generation of potent broadly neutralizing HIV-1 antibodies (bnAbs) has generated substantial interest in their potential use for the prevention and/or treatment of HIV-1 infection. While combinations of bnAbs targeting distinct epitopes on the viral envelope (Env) will likely be required to overcome the extraordinary diversity of HIV-1, a key outstanding question is which bnAbs, and how many, will be needed to achieve optimal clinical benefit. We assessed the neutralizing activity of 15 bnAbs targeting four distinct epitopes of Env, including the CD4-binding site (CD4bs), the V1/V2-glycan region, the V3-glycan region, and the gp41 membrane proximal external region (MPER), against a panel of 200 acute/early clade C HIV-1 Env pseudoviruses. A mathematical model was developed that predicted neutralization by a subset of experimentally evaluated bnAb combinations with high accuracy. Using this model, we performed a comprehensive and systematic comparison of the predicted neutralizing activity of over 1,600 possible double, triple, and quadruple bnAb combinations. The most promising bnAb combinations were identified based not only on breadth and potency of neutralization, but also other relevant measures, such as the extent of complete neutralization and instantaneous inhibitory potential (IIP). By this set of criteria, triple and quadruple combinations of bnAbs were identified that were significantly more effective than the best double combinations, and further improved the probability of having multiple bnAbs simultaneously active against a given virus, a requirement that may be critical for countering escape in vivo. These results provide a rationale for advancing bnAb combinations with the best in vitro predictors of success into clinical trials for both the prevention and treatment of HIV-1 infection.