Genetic association analysis of human median voice pitch identifies a common locus for tonal and non-tonal languages.

The genetic influence on human vocal pitch in tonal and non-tonal languages remains largely unknown. In tonal languages, such as Mandarin Chinese, pitch changes differentiate word meanings, whereas in non-tonal languages, such as Icelandic, pitch is used to convey intonation. We addressed this question by searching for genetic associations with interindividual variation in median pitch in a Chinese major depression case-control cohort and compared our results with a genome-wide association study from Iceland. The same genetic variant, rs11046212-T in an intron of the ABCC9 gene, was one of the most strongly associated loci with median pitch in both samples. Our meta-analysis revealed four genome-wide significant hits, including two novel associations. The discovery of genetic variants influencing vocal pitch across both tonal and non-tonal languages suggests the possibility of a common genetic contribution to the human vocal system shared in two distinct populations with languages that differ in tonality (Icelandic and Mandarin).

Complementation testing identifies genes mediating effects at quantitative trait loci underlying fear-related behavior.

Knowing the genes involved in quantitative traits provides an entry point to understanding the biological bases of behavior, but there are very few examples where the pathway from genetic locus to behavioral change is known. To explore the role of specific genes in fear behavior, we mapped three fear-related traits, tested fourteen genes at six quantitative trait loci (QTLs) by quantitative complementation, and identified six genes. Four genes, Lamp, Ptprd, Nptx2, and Sh3gl, have known roles in synapse function; the fifth, Psip1, was not previously implicated in behavior; and the sixth is a long non-coding RNA, 4933413L06Rik, of unknown function. Variation in transcriptome and epigenetic modalities occurred preferentially in excitatory neurons, suggesting that genetic variation is more permissible in excitatory than inhibitory neuronal circuits. Our results relieve a bottleneck in using genetic mapping of QTLs to uncover biology underlying behavior and prompt a reconsideration of expected relationships between genetic and functional variation.

A Privacy-Preserving Unsupervised Speaker Disentanglement Method for Depression Detection from Speech.

The proposed method focuses on speaker disentanglement in the context of depression detection from speech signals. Previous approaches require patient/speaker labels, encounter instability due to loss maximization, and introduce unnecessary parameters for adversarial domain prediction. In contrast, the proposed unsupervised approach reduces cosine similarity between latent spaces of depression and pre-trained speaker classification models. This method outperforms baseline models, matches or exceeds adversarial methods in performance, and does so without relying on speaker labels or introducing additional model parameters, leading to a reduction in model complexity. The higher the speaker de-identification score (DeID), the better the depression detection system is in masking a patient's identity thereby enhancing the privacy attributes of depression detection systems. On the DAIC-WOZ dataset with ComparE16 features and an LSTM-only model, our method achieves an F1-Score of 0.776 and a DeID score of 92.87%, outperforming its adversarial counterpart which has an F1Score of 0.762 and 68.37% DeID, respectively. Furthermore, we demonstrate that speaker-disentanglement methods are complementary to text-based approaches, and a score-level fusion with a Word2vec-based depression detection model further enhances the overall performance to an F1-Score of 0.830.

Enhancing accuracy and privacy in speech-based depression detection through speaker disentanglement.

Speech signals are valuable biomarkers for assessing an individual's mental health, including identifying Major Depressive Disorder (MDD) automatically. A frequently used approach in this regard is to employ features related to speaker identity, such as speaker-embeddings. However, over-reliance on speaker identity features in mental health screening systems can compromise patient privacy. Moreover, some aspects of speaker identity may not be relevant for depression detection and could serve as a bias factor that hampers system performance. To overcome these limitations, we propose disentangling speaker-identity information from depression-related information. Specifically, we present four distinct disentanglement methods to achieve this - adversarial speaker identification (SID)-loss maximization (ADV), SID-loss equalization with variance (LEV), SID-loss equalization using Cross-Entropy (LECE) and SID-loss equalization using KL divergence (LEKLD). Our experiments, which incorporated diverse input features and model architectures, have yielded improved F1 scores for MDD detection and voice-privacy attributes, as quantified by Gain in Voice Distinctiveness GV D and De-Identification Scores (DeID). On the DAIC-WOZ dataset (English), LECE using ComparE16 features results in the best F1-Scores of 80% which represents the audio-only SOTA depression detection F1-Score along with a GV D of -1.1 dB and a DeID of 85%. On the EATD dataset (Mandarin), ADV using raw-audio signal achieves an F1-Score of 72.38% surpassing multi-modal SOTA along with a GV D of -0.89 dB dB and a DeID of 51.21%. By reducing the dependence on speaker-identity-related features, our method offers a promising direction for speech-based depression detection that preserves patient privacy.

Personalized mood prediction from patterns of behavior collected with smartphones.

Over the last ten years, there has been considerable progress in using digital behavioral phenotypes, captured passively and continuously from smartphones and wearable devices, to infer depressive mood. However, most digital phenotype studies suffer from poor replicability, often fail to detect clinically relevant events, and use measures of depression that are not validated or suitable for collecting large and longitudinal data. Here, we report high-quality longitudinal validated assessments of depressive mood from computerized adaptive testing paired with continuous digital assessments of behavior from smartphone sensors for up to 40 weeks on 183 individuals experiencing mild to severe symptoms of depression. We apply a combination of cubic spline interpolation and idiographic models to generate individualized predictions of future mood from the digital behavioral phenotypes, achieving high prediction accuracy of depression severity up to three weeks in advance (R2 ≥ 80%) and a 65.7% reduction in the prediction error over a baseline model which predicts future mood based on past depression severity alone. Finally, our study verified the feasibility of obtaining high-quality longitudinal assessments of mood from a clinical population and predicting symptom severity weeks in advance using passively collected digital behavioral data. Our results indicate the possibility of expanding the repertoire of patient-specific behavioral measures to enable future psychiatric research.

Complementation testing identifies causal genes at quantitative trait loci underlying fear related behavior.

Knowing the genes involved in quantitative traits provides a critical entry point to understanding the biological bases of behavior, but there are very few examples where the pathway from genetic locus to behavioral change is known. Here we address a key step towards that goal by deploying a test that directly queries whether a gene mediates the effect of a quantitative trait locus (QTL). To explore the role of specific genes in fear behavior, we mapped three fear-related traits, tested fourteen genes at six QTLs, and identified six genes. Four genes, Lsamp, Ptprd, Nptx2 and Sh3gl, have known roles in synapse function; the fifth gene, Psip1, is a transcriptional co-activator not previously implicated in behavior; the sixth is a long non-coding RNA 4933413L06Rik with no known function. Single nucleus transcriptomic and epigenetic analyses implicated excitatory neurons as likely mediating the genetic effects. Surprisingly, variation in transcriptome and epigenetic modalities between inbred strains occurred preferentially in excitatory neurons, suggesting that genetic variation is more permissible in excitatory than inhibitory neuronal circuits. Our results open a bottleneck in using genetic mapping of QTLs to find novel biology underlying behavior and prompt a reconsideration of expected relationships between genetic and functional variation.

An evolutionary perspective on complex neuropsychiatric disease.

The forces of evolution-mutation, selection, migration, and genetic drift-shape the genetic architecture of human traits, including the genetic architecture of complex neuropsychiatric illnesses. Studying these illnesses in populations that are diverse in genetic ancestry, historical demography, and cultural history can reveal how evolutionary forces have guided adaptation over time and place. A fundamental truth of shared human biology is that an allele responsible for a disease in anyone, anywhere, reveals a gene critical to the normal biology underlying that condition in everyone, everywhere. Understanding the genetic causes of neuropsychiatric disease in the widest possible range of human populations thus yields the greatest possible range of insight into genes critical to human brain development. In this perspective, we explore some of the relationships between genes, adaptation, and history that can be illuminated by an evolutionary perspective on studies of complex neuropsychiatric disease in diverse populations.

Single-cell methylation analysis of brain tissue prioritizes mutations that alter transcription.

Relating genetic variants to behavior remains a fundamental challenge. To assess the utility of DNA methylation marks in discovering causative variants, we examined their relationship to genetic variation by generating single-nucleus methylomes from the hippocampus of eight inbred mouse strains. At CpG sequence densities under 40 CpG/Kb, cells compensate for loss of methylated sites by methylating additional sites to maintain methylation levels. At higher CpG sequence densities, the exact location of a methylated site becomes more important, suggesting that variants affecting methylation will have a greater effect when occurring in higher CpG densities than in lower. We found this to be true for a variant's effect on transcript abundance, indicating that candidate variants can be prioritized based on CpG sequence density. Our findings imply that DNA methylation influences the likelihood that mutations occur at specific sites in the genome, supporting the view that the distribution of mutations is not random.

Deep learning-based phenotype imputation on population-scale biobank data increases genetic discoveries.

Biobanks that collect deep phenotypic and genomic data across many individuals have emerged as a key resource in human genetics. However, phenotypes in biobanks are often missing across many individuals, limiting their utility. We propose AutoComplete, a deep learning-based imputation method to impute or 'fill-in' missing phenotypes in population-scale biobank datasets. When applied to collections of phenotypes measured across ~300,000 individuals from the UK Biobank, AutoComplete substantially improved imputation accuracy over existing methods. On three traits with notable amounts of missingness, we show that AutoComplete yields imputed phenotypes that are genetically similar to the originally observed phenotypes while increasing the effective sample size by about twofold on average. Further, genome-wide association analyses on the resulting imputed phenotypes led to a substantial increase in the number of associated loci. Our results demonstrate the utility of deep learning-based phenotype imputation to increase power for genetic discoveries in existing biobank datasets.

Phenotype integration improves power and preserves specificity in biobank-based genetic studies of major depressive disorder.

Biobanks often contain several phenotypes relevant to diseases such as major depressive disorder (MDD), with partly distinct genetic architectures. Researchers face complex tradeoffs between shallow (large sample size, low specificity/sensitivity) and deep (small sample size, high specificity/sensitivity) phenotypes, and the optimal choices are often unclear. Here we propose to integrate these phenotypes to combine the benefits of each. We use phenotype imputation to integrate information across hundreds of MDD-relevant phenotypes, which significantly increases genome-wide association study (GWAS) power and polygenic risk score (PRS) prediction accuracy of the deepest available MDD phenotype in UK Biobank, LifetimeMDD. We demonstrate that imputation preserves specificity in its genetic architecture using a novel PRS-based pleiotropy metric. We further find that integration via summary statistics also enhances GWAS power and PRS predictions, but can introduce nonspecific genetic effects depending on input. Our work provides a simple and scalable approach to improve genetic studies in large biobanks by integrating shallow and deep phenotypes.

Dominance is common in mammals and is associated with trans-acting gene expression and alternative splicing.

BACKGROUND: Dominance and other non-additive genetic effects arise from the interaction between alleles, and historically these phenomena play a major role in quantitative genetics. However, most genome-wide association studies (GWAS) assume alleles act additively. RESULTS: We systematically investigate both dominance-here representing any non-additive within-locus interaction-and additivity across 574 physiological and gene expression traits in three mammalian stocks: F2 intercross pigs, rat heterogeneous stock, and mice heterogeneous stock. Dominance accounts for about one quarter of heritable variance across all physiological traits in all species. Hematological and immunological traits exhibit the highest dominance variance, possibly reflecting balancing selection in response to pathogens. Although most quantitative trait loci (QTLs) are detectable as additive QTLs, we identify 154, 64, and 62 novel dominance QTLs in pigs, rats, and mice respectively that are undetectable as additive QTLs. Similarly, even though most cis-acting expression QTLs are additive, gene expression exhibits a large fraction of dominance variance, and trans-acting eQTLs are enriched for dominance. Genes causal for dominance physiological QTLs are less likely to be physically linked to their QTLs but instead act via trans-acting dominance eQTLs. In addition, thousands of eQTLs are associated with alternatively spliced isoforms with complex additive and dominant architectures in heterogeneous stock rats, suggesting a possible mechanism for dominance. CONCLUSIONS: Although heritability is predominantly additive, many mammalian genetic effects are dominant and likely arise through distinct mechanisms. It is therefore advantageous to consider both additive and dominance effects in GWAS to improve power and uncover causality.

Phenotypic subtyping via contrastive learning.

Defining and accounting for subphenotypic structure has the potential to increase statistical power and provide a deeper understanding of the heterogeneity in the molecular basis of complex disease. Existing phenotype subtyping methods primarily rely on clinically observed heterogeneity or metadata clustering. However, they generally tend to capture the dominant sources of variation in the data, which often originate from variation that is not descriptive of the mechanistic heterogeneity of the phenotype of interest; in fact, such dominant sources of variation, such as population structure or technical variation, are, in general, expected to be independent of subphenotypic structure. We instead aim to find a subspace with signal that is unique to a group of samples for which we believe that subphenotypic variation exists (e.g., cases of a disease). To that end, we introduce Phenotype Aware Components Analysis (PACA), a contrastive learning approach leveraging canonical correlation analysis to robustly capture weak sources of subphenotypic variation. In the context of disease, PACA learns a gradient of variation unique to cases in a given dataset, while leveraging control samples for accounting for variation and imbalances of biological and technical confounders between cases and controls. We evaluated PACA using an extensive simulation study, as well as on various subtyping tasks using genotypes, transcriptomics, and DNA methylation data. Our results provide multiple strong evidence that PACA allows us to robustly capture weak unknown variation of interest while being calibrated and well-powered, far superseding the performance of alternative methods. This renders PACA as a state-of-the-art tool for defining de novo subtypes that are more likely to reflect molecular heterogeneity, especially in challenging cases where the phenotypic heterogeneity may be masked by a myriad of strong unrelated effects in the data.

The genetic basis of major depressive disorder.

The genetic dissection of major depressive disorder (MDD) ranks as one of the success stories of psychiatric genetics, with genome-wide association studies (GWAS) identifying 178 genetic risk loci and proposing more than 200 candidate genes. However, the GWAS results derive from the analysis of cohorts in which most cases are diagnosed by minimal phenotyping, a method that has low specificity. I review data indicating that there is a large genetic component unique to MDD that remains inaccessible to minimal phenotyping strategies and that the majority of genetic risk loci identified with minimal phenotyping approaches are unlikely to be MDD risk loci. I show that inventive uses of biobank data, novel imputation methods, combined with more interviewer diagnosed cases, can identify loci that contribute to the episodic severe shifts of mood, and neurovegetative and cognitive changes that are central to MDD. Furthermore, new theories about the nature and causes of MDD, drawing upon advances in neuroscience and psychology, can provide handles on how best to interpret and exploit genetic mapping results.

Unsupervised Instance Discriminative Learning for Depression Detection from Speech Signals.

Major Depressive Disorder (MDD) is a severe illness that affects millions of people, and it is critical to diagnose this disorder as early as possible. Detecting depression from voice signals can be of great help to physicians and can be done without any invasive procedure. Since relevant labelled data are scarce, we propose a modified Instance Discriminative Learning (IDL) method, an unsupervised pre-training technique, to extract augment-invariant and instance-spread-out embeddings. In terms of learning augment-invariant embeddings, various data augmentation methods for speech are investigated, and time-masking yields the best performance. To learn instance-spreadout embeddings, we explore methods for sampling instances for a training batch (distinct speaker-based and random sampling). It is found that the distinct speaker-based sampling provides better performance than the random one, and we hypothesize that this result is because relevant speaker information is preserved in the embedding. Additionally, we propose a novel sampling strategy, Pseudo Instance-based Sampling (PIS), based on clustering algorithms, to enhance spread-out characteristics of the embeddings. Experiments are conducted with DepAudioNet on DAIC-WOZ (English) and CONVERGE (Mandarin) datasets, and statistically significant improvements, with p-value 0.0015 and 0.05, respectively, are observed using PIS in the detection of MDD relative to the baseline without pre-training.

A Step Towards Preserving Speakers' Identity While Detecting Depression Via Speaker Disentanglement.

Preserving a patient's identity is a challenge for automatic, speech-based diagnosis of mental health disorders. In this paper, we address this issue by proposing adversarial disentanglement of depression characteristics and speaker identity. The model used for depression classification is trained in a speaker-identity-invariant manner by minimizing depression prediction loss and maximizing speaker prediction loss during training. The effectiveness of the proposed method is demonstrated on two datasets - DAIC-WOZ (English) and CONVERGE (Mandarin), with three feature sets (Mel-spectrograms, raw-audio signals, and the last-hidden-state of Wav2vec2.0), using a modified DepAudioNet model. With adversarial training, depression classification improves for every feature when compared to the baseline. Wav2vec2.0 features with adversarial learning resulted in the best performance (F1-score of 69.2% for DAIC-WOZ and 91.5% for CONVERGE). Analysis of the class-separability measure (J-ratio) of the hidden states of the DepAudioNet model shows that when adversarial learning is applied, the backend model loses some speaker-discriminability while it improves depression-discriminability. These results indicate that there are some components of speaker identity that may not be useful for depression detection and minimizing their effects provides a more accurate diagnosis of the underlying disorder and can safeguard a speaker's identity.

Cross-trait assortative mating is widespread and inflates genetic correlation estimates.

The observation of genetic correlations between disparate human traits has been interpreted as evidence of widespread pleiotropy. Here, we introduce cross-trait assortative mating (xAM) as an alternative explanation. We observe that xAM affects many phenotypes and that phenotypic cross-mate correlation estimates are strongly associated with genetic correlation estimates (R2=74%). We demonstrate that existing xAM plausibly accounts for substantial fractions of genetic correlation estimates and that previously reported genetic correlation estimates between some pairs of psychiatric disorders are congruent with xAM alone. Finally, we provide evidence for a history of xAM at the genetic level using cross-trait even/odd chromosome polygenic score correlations. Together, our results demonstrate that previous reports have likely overestimated the true genetic similarity between many phenotypes.

Identification of genetic mechanisms for tissue-specific genetic effects based on CRISPR screens.

A major challenge in genetic studies of complex diseases is to determine how the action of risk genes is restricted to a tissue or cell type. Here, we investigate tissue specificity of gene action using CRISPR screens from 786 cancer cell lines originating from 24 tissues. We find that the expression pattern of the gene across tissues explains only a minority of cases of tissue-specificity (9%), while gene amplification and the expression levels of paralogs account for 39.5% and 15.5%, respectively. In addition, the transfer of small molecules to mutant cells explains tissue-specific gene action in blood. The tissue-specific genes we found are not specific just for human cancer cell lines: we found that the tissue-specific genes are intolerant to functional mutations in the human population and are associated with human diseases more than genes that are essential across all cell types. Our findings offer important insights into genetic mechanisms for tissue specificity of human diseases.

A comprehensive benchmarking of WGS-based deletion structural variant callers.

Advances in whole-genome sequencing (WGS) promise to enable the accurate and comprehensive structural variant (SV) discovery. Dissecting SVs from WGS data presents a substantial number of challenges and a plethora of SV detection methods have been developed. Currently, evidence that investigators can use to select appropriate SV detection tools is lacking. In this article, we have evaluated the performance of SV detection tools on mouse and human WGS data using a comprehensive polymerase chain reaction-confirmed gold standard set of SVs and the genome-in-a-bottle variant set, respectively. In contrast to the previous benchmarking studies, our gold standard dataset included a complete set of SVs allowing us to report both precision and sensitivity rates of the SV detection methods. Our study investigates the ability of the methods to detect deletions, thus providing an optimistic estimate of SV detection performance as the SV detection methods that fail to detect deletions are likely to miss more complex SVs. We found that SV detection tools varied widely in their performance, with several methods providing a good balance between sensitivity and precision. Additionally, we have determined the SV callers best suited for low- and ultralow-pass sequencing data as well as for different deletion length categories.

FRAUG: A FRAME RATE BASED DATA AUGMENTATION METHOD FOR DEPRESSION DETECTION FROM SPEECH SIGNALS.

In this paper, a data augmentation method is proposed for depression detection from speech signals. Samples for data augmentation were created by changing the frame-width and the frame-shift parameters during the feature extraction process. Unlike other data augmentation methods (such as VTLP, pitch perturbation, or speed perturbation), the proposed method does not explicitly change acoustic parameters but rather the time-frequency resolution of frame-level features. The proposed method was evaluated using two different datasets, models, and input acoustic features. For the DAIC-WOZ (English) dataset when using the DepAudioNet model and mel-Spectrograms as input, the proposed method resulted in an improvement of 5.97% (validation) and 25.13% (test) when compared to the baseline. The improvements for the CONVERGE (Mandarin) dataset when using the x-vector embeddings with CNN as the backend and MFCCs as input features were 9.32% (validation) and 12.99% (test). Baseline systems do not incorporate any data augmentation. Further, the proposed method outperformed commonly used data-augmentation methods such as noise augmentation, VTLP, Speed, and Pitch Perturbation. All improvements were statistically significant.

Genomic epidemiology of the Los Angeles COVID-19 outbreak and the early history of the B.1.43 strain in the USA.

BACKGROUND: The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused global disruption of human health and activity. Being able to trace the early outbreak of SARS-CoV-2 within a locality can inform public health measures and provide insights to contain or prevent viral transmission. Investigation of the transmission history requires efficient sequencing methods and analytic strategies, which can be generally useful in the study of viral outbreaks. METHODS: The County of Los Angeles (hereafter, LA County) sustained a large outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). To learn about the transmission history, we carried out surveillance viral genome sequencing to determine 142 viral genomes from unique patients seeking care at the University of California, Los Angeles (UCLA) Health System. 86 of these genomes were from samples collected before April 19, 2020. RESULTS: We found that the early outbreak in LA County, as in other international air travel hubs, was seeded by multiple introductions of strains from Asia and Europe. We identified a USA-specific strain, B.1.43, which was found predominantly in California and Washington State. While samples from LA County carried the ancestral B.1.43 genome, viral genomes from neighboring counties in California and from counties in Washington State carried additional mutations, suggesting a potential origin of B.1.43 in Southern California. We quantified the transmission rate of SARS-CoV-2 over time, and found evidence that the public health measures put in place in LA County to control the virus were effective at preventing transmission, but might have been undermined by the many introductions of SARS-CoV-2 into the region. CONCLUSION: Our work demonstrates that genome sequencing can be a powerful tool for investigating outbreaks and informing the public health response. Our results reinforce the critical need for the USA to have coordinated inter-state responses to the pandemic.