Biology subjects

de los Campos, G.

Publications and source records attributed to de los Campos, G..

13 recordsLinked to original sources

bioRxiv2024

Global Genotype by Environment Prediction Competition Reveals That Diverse Modeling Strategies Can Deliver Satisfactory Maize Yield Estimates

Predicting phenotypes from a combination of genetic and environmental factors is a grand challenge of modern biology. Slight improvements in this area have the potential to save lives, improve food and fuel security, permit better care of the planet, and create other positive outcomes. In 2022 and 2023 the first open-to-the-public Genomes to Fields (G2F) initiative Genotype by Environment (GxE) prediction competition was held using a large dataset including genomic variation, phenotype and weather measurements and field management notes, gathered by the project over nine years. The competition attracted registrants from around the world with representation from academic, government, industry, and non-profit institutions as well as unaffiliated. These participants came from diverse disciplines include plant science, animal science, breeding, statistics, computational biology and others. Some participants had no formal genetics or plant-related training, and some were just beginning their graduate education. The teams applied varied methods and strategies, providing a wealth of modeling knowledge based on a common dataset. The winners strategy involved two models combining machine learning and traditional breeding tools: one model emphasized environment using features extracted by Random Forest, Ridge Regression and Least-squares, and one focused on genetics. Other high-performing teams methods included quantitative genetics, classical machine learning/deep learning, mechanistic models, and model ensembles. The dataset factors used, such as genetics; weather; and management data, were also diverse, demonstrating that no single model or strategy is far superior to all others within the context of this competition.

genetics↗

bioRxiv2024

A conserved phenylalanine motif among Teleost fish provides insight for improving electromagnetic perception.

Magnetoreceptive biology as a field remains relatively obscure; compared to the breadth of species believed to sense magnetic fields, it remains under-studied. Here, we present grounds for the expansion of magnetoreception studies among Teleosts. We begin with the electromagnetic perceptive gene (EPG) from Kryptopterus vitreolus and expand to identify 72 Teleosts with homologous proteins containing a conserved three-phenylalanine (3F) motif. Phylogenetic analysis provides insight as to how EPG may have evolved over time, and indicates that certain clades may have experienced a loss of function driven by different fitness pressures. One potential factor is water type with freshwater fish significantly more likely to possess the functional motif version (FFF), and saltwater fish to have the non-functional variant (FXF). It was also revealed that when the 3F motif from the homolog of Brachyhypopomus gauderio (B.g.) is inserted into EPG - EPG(B.g.) - the response (as indicated by increased intracellular calcium) is faster. This indicates that EPG has the potential to be engineered to improve upon its response and increase its utility to be used as a controller for specific outcomes.

synthetic biology↗

bioRxiv2024

Bayesian Hierarchical Hypothesis Testing in Large-Scale Genome-Wide Association Analysis

Variable selection and large-scale hypothesis testing are techniques commonly used to analyze high-dimensional genomic data. Despite recent advances in theory and methodology, variable selection and inference with highly collinear features remain challenging. For instance, collinearity poses a great challenge in Genome-Wide Association Studies (GWAS) involving millions of variants, many of which may be in high linkage disequilibrium. In such settings, collinearity can significantly reduce the power of variable selection methods to identify individual variants associated with an outcome. To address such challenges, we developed a Bayesian Hierarchical Hypothesis Testing (BHHT)-a novel multi-resolution testing procedure that offers high power with adequate error control and fine-mapping resolution. We demonstrate through simulations that the proposed methodology has a power-FDR performance that is competitive with (and in many scenarios better than) state-of-the-art methods. Finally, we demonstrate the feasibility of using the proposed methodology with big data to map risk variants for serum urate using data (n[~]300,000) on phenotype and ultra-high-dimensional genotypes ([~]15 million SNPs) from the UK-Biobank. Our results show that the proposed methodology leads to many more discoveries than those obtained using traditional feature-centered inference procedures. The article is accompanied by open-source software that implements the methods described in this study using algorithms that scale to biobank-size ultra-high-dimensional data.

genetics↗

bioRxiv2020

Opportunities and limits of combining microbiome and genome data for complex trait prediction

The analysis and prediction of complex traits using microbiome data combined with host genomic information is a topic of utmost interest. However, numerous questions remain to be answered: How useful can the microbiome be for complex trait prediction? Are microbiability estimates reliable? Can the underlying biological links between the hosts genome, microbiome, and the phenome be recovered? Here, we address these issues by (i) developing a novel simulation strategy that uses real microbiome and genotype data as input, and (ii) proposing a variance-component approach which, in the spirit of mediation analyses, quantifies the proportion of phenotypic variance explained by genome and microbiome, and dissects it into direct and indirect effects. The proposed simulation approach can mimic a genetic link between the microbiome and SNP data via a permutation procedure that retains the distributional properties of the data. Results suggest that microbiome data could significantly improve phenotype prediction accuracy, irrespective of whether some abundances are under direct genetic control by the host or not. Overall, random-effects linear methods appear robust for variance components estimation, despite the highly leptokurtic distribution of microbiota abundances. Nevertheless, we observed that accuracy depends in part on the number of microorganisms taxa influencing the trait of interest. While we conclude that overall genome-microbiome-links can be characterized via variance components, we are less optimistic about the possibility of identifying the causative effects, i.e., individual SNPs affecting abundances; power at this level would require much larger sample sizes than the ones typically available for genome-microbiome-phenome data. Author summaryThe microbiome consists of the microorganisms that live in a particular environment, including those in our organism. There is consistent evidence that these communities play an important role in numerous traits of relevance, including disease susceptibility or feed efficiency. Moreover, it has been shown that the microbiome can be relatively stable throughout an individuals life and that is affected by the host genome. These reasons have prompted numerous studies to determine whether and how the microbiome can be used for prediction of complex phenotypes, either using microbiome alone or in combination with hosts genome data. However, numerous questions remain to be answered such as the reliability of parameter estimates, or which is the underlying relationship between microbiome, genome, and phenotype. The few available empirical studies do not provide a clear answer to these problems. Here we address these issues by developing a novel simulation strategy and we show that, although the microbiome can significantly help in prediction, it will be difficult to retrieve the actual biological basis of interactions between the microbiome and the trait.

genetics↗

bioRxiv2019

Microarray Gene Expression DatasetRe-Analysis Reveals Variability in InfluenzaInfection and Vaccination.

Influenza, a communicable disease, affects thousands of people worldwide. Young children, elderly, immunocompromised individuals and pregnant women are at higher risk for being infected by the influenza virus. Our study aims to highlight differentially expressed genes in influenza disease compared to influenza vaccination. We also investigate genetic variation due to the age and sex of samples. To accomplish our goals, we conducted a meta-analysis using publicly available microarray expression data. Our inclusion criteria included subjects with influenza, subjects who received the influenza vaccine and healthy controls. We curated 18 microarray datasets for a total of 3,481 samples (1,277 controls, 297 influenza infection, 1,907 influenza vaccination). We pre-processed the raw microarray expression data in R using packages available to pre-process Affymetrix and Illumina microarray platforms. We used a Box-Cox power transformation of the data prior to our down-stream analysis to identify differentially expressed genes. Statistical analyses were based on linear mixed effects model with all study factors and successive likelihood ratio tests (LRT) to identify differentially-expressed genes. We filtered LRT results by disease (Bonferroni adjusted p-value < 0.05) and used a two-tailed 10% quantile cutoff to identify biologically significant genes. Furthermore, we assessed age and sex effects on the disease genes by filtering for genes with a statistically significant (Bonferroni adjusted p-value < 0.05) interaction between disease and age, and disease and sex. We identified 4,889 statistically significant genes when we filtered the LRT results by disease factor, and gene enrichment analysis (gene ontology and pathways) included innate immune response, viral process, defense response to virus, Hematopoietic cell lineage and NF-kappa B signaling pathway. Our quantile filtered gene lists comprised of 978 genes each associated with influenza infection and vaccination. We also identified 907 and 48 genes with statistically significant (Bonferroni adjusted p-value < 0.05) disease-age and disease-sex interactions respectively. Our meta-analysis approach highlights key gene signatures and their associated pathways for both influenza infection and vaccination. We also were able to identify genes with an age and sex effect. This gives potential for improving current vaccines and exploring genes that are expressed equally across ages when considering universal vaccinations for influenza.

immunology↗

bioRxiv2019

Deciphering sex-specific genetic architectures using local Bayesian regressions

Many complex human traits exhibit differences between sexes. While numerous factors likely contribute to this phenomenon, growing evidence from genome-wide studies suggest a partial explanation: that males and females from the same population possess differing genetic architectures. Despite this, mapping gene-by-sex (GxS) interactions remains a challenge likely because the magnitude of such an interaction is typically and exceedingly small; traditional genome-wide association techniques may be underpowered to detect such events partly due to the burden of multiple test correction. Here, we developed a local Bayesian regression (LBR) method to estimate sex-specific SNP marker effects after fully accounting for local linkage-disequilibrium (LD) patterns. This enabled us to infer sex-specific effects and GxS interactions either at the single SNP level, or by aggregating the effects of multiple SNPs to make inferences at the level of small LD-based regions. Using simulations in which there was imperfect LD between SNPs and causal variants, we showed that aggregating sex-specific marker effects with LBR provides improved power and resolution to detect GxS interactions over traditional single-SNP-based tests. When using LBR to analyze traits from the UK Biobank, we detected a relatively large GxS interaction impacting bone-mineral density within ABO and replicated many previously detected large-magnitude GxS interactions impacting waist-to-hip ratio. We also discovered many new GxS interactions impacting such traits as height and BMI within regions of the genome where both male- and female-specific effects explain a small proportion of phenotypic variance (R2 < 1x10-4), but are enriched in known expression quantitative trait loci. By combining biobank-level data and techniques to estimate sex-specific SNP effects after accounting for local-LD patterns, we are providing evidence that numerous small-magnitude GxS interactions exist to influence sex differences in a variety of complex traits.\n\nAuthor SummaryMany complex human traits are known to be influenced by an impressive number of causal variants each with very small effects, posing great challenges for genome-wide association studies (GWAS). To add to this challenge, many causal variants may possess context-dependent effects such as effects that are dependent on biological sex. While GWAS are commonly performed using specific methods in which one single nucleotide polymorphism (SNP) at a time is tested for association with a trait, alternatively we utilize methods more commonly observed in the genomic prediction literature. Such methods are advantageous in that they are not burdened by multiple test correction in the same way as traditional GWAS techniques are, and can fully account for linkage-disequilibrium patterns to accurately estimate the true effects of SNP markers. Here we adapt such methods to estimate genetic effects within sexes and provide a powerful means to compare sex-specific genetic effects.

genetics↗

bioRxiv2019

Genetic image-processing using regularized selection indices

High-throughput phenotyping (HTP) technologies can produce data on thousands of phenotypes per unit being monitored. These data can be used to breed for economically and environmentally relevant traits (e.g., drought tolerance); however, incorporating high-dimensional phenotypes in genetic analyses and in breeding schemes poses important statistical and computational challenges. To address this problem, we developed regularized selection indices; the methodology integrates techniques commonly used in high-dimensional phenotypic regressions (including penalization and rank-reduction approaches) into the selection index (SI) framework. Using extensive data from CIMMYTs (International Maize and Wheat Improvement Center) wheat breeding program we show that regularized SIs derived from hyper-spectral data offer consistently higher accuracy for grain yield than those achieved by canonical SIs, and by vegetation indices commonly used to predict agronomic traits. Regularized SIs offer an effective approach to leverage HTP data that is routinely generated in agriculture; the methodology can also be used to conduct genetic studies using high-dimensional phenotypes that are often collected in humans and model organisms including body images and whole-genome gene expression profiles.

genetics↗

bioRxiv2019

Benchmarking algorithms for genomic prediction of complex traits

The usefulness of Genomic Prediction (GP) in crop and livestock breeding programs has led to efforts to develop new and improved GP approaches including non-linear algorithm, such as artificial neural networks (ANN) (i.e. deep learning) and gradient tree boosting. However, the performance of these algorithms has not been compared in a systematic manner using a wide range of GP datasets and models. Using data of 18 traits across six plant species with different marker densities and training population sizes, we compared the performance of six linear and five non-linear algorithms, including ANNs. First, we found that hyperparameter selection was critical for all non-linear algorithms and that feature selection prior to model training was necessary for ANNs when the markers greatly outnumbered the number of training lines. Across all species and trait combinations, no one algorithm performed best, however predictions based on a combination of results from multiple GP algorithms (i.e. ensemble predictions) performed consistently well. While linear and non-linear algorithms performed best for a similar number of traits, the performance of non-linear algorithms vary more between traits than that of linear algorithms. Although ANNs did not perform best for any trait, we identified strategies (i.e. feature selection, seeded starting weights) that boosted their performance near the level of other algorithms. These results, together with the fact that even small improvements in GP performance could accumulate into large genetic gains over the course of a breeding program, highlights the importance of algorithm selection for the prediction of trait values.

genetics↗

bioRxiv2019

Transcriptome-based prediction of complex traits in maize

The ability to predict traits from genome-wide sequence information (Genomic Prediction, GP), has improved our understanding of the genetic basis of complex traits and transformed breeding practices. Transcriptome data may also be useful for GP. However, it remains unclear how well transcript levels can predict traits, particularly when traits are scored at different development stages. Using maize genetic markers and transcript levels from seedlings to predict mature plant traits, we found transcript and genetic marker models have similar performance. Surprisingly, genetic markers important for predictions were not close to or identified as regulatory variants for important transcripts. Thus, transcript levels are predictive not simply due to genetic variation. Furthermore, genetic marker models identified only one of 14 benchmark flowering time genes, while transcript models identified five. Our findings highlight that transcriptome data is useful for GP and can provide a link between traits and variation that cannot be readily captured at the sequence level.

genetics↗

bioRxiv2018

Quantifying Heterogeneity in the Genetic Architecture of Complex Traits Between Ethnically Diverse Groups using Random Effect Interaction Models

In humans, most genome-wide association studies have been conducted using data from Caucasians and many of the reported findings have not replicated in other populations. This lack of replication may be due to statistical issues (small sample size, confounding) or perhaps more fundamentally to differences in the genetic architecture of traits between ethnically diverse subpopulations. What aspects of the genetic architecture of traits vary between subpopulations and how can this be quantified? We consider studying effect heterogeneity using random-effect Bayesian interaction models. The proposed methodology can be applied using shrinkage and variable selection methods and produces useful information about effect heterogeneity in the form of whole-genome summaries (e.g., SNP-heritability and the average correlation of effects) as well as SNP-specific attributes. Using simulations, we show that the proposed methodology yields (nearly) unbiased estimates of genomic heritability and of the average correlation of effects between groups when the sample size is not too small relative to the number of SNPs used. Subsequently, we used the proposed methodology for the analyses of four complex human traits (standing height, high-density lipoprotein, low-density lipoprotein, and serum urate levels) in European-Americans (EAs) and African-Americans (AAs). The estimated correlations of effects between the two subpopulations was well below unity for all the traits, ranging from 0.73 to 0.50. The extent of effect heterogeneity varied between traits and SNP-sets. Height showed less differences in SNP effects between AAs and EAs whereas HDL, a trait highly influenced by life-style, exhibited greater extent of effect heterogeneity. For all the traits we observed substantial variability in effect heterogeneity across SNPs, suggesting it varies between regions of the genome.

genetics↗

bioRxiv2018

Imperfect Linkage Disequilibrium Generates Phantom Epistasis (& Perils of Big Data)

The genetic architecture of complex human traits and diseases is affected by large number of possibly interacting genes, but detecting epistatic interactions can be challenging. In the last decade, several studies have alluded to problems that linkage disequilibrium can create when testing for epistatic interactions between DNA markers. However, these problems have not been formalized nor have their consequences been quantified in a precise manner. Here we use a conceptually simple three locus model involving a causal locus and two markers to show that imperfect LD can generate the illusion of epistasis, even when the underlying genetic architecture is purely additive. We describe necessary conditions for such \"phantom epistasis\" to emerge and quantify its relevance using simulations. Our empirical results demonstrate that phantom epistasis can be a very serious problem in GWAS studies (with rejection rates against the additive model greater than 0.2 for nominal p-values of 0.05, even when the model is purely additive). Some studies have sought to avoid this problem by only testing interactions between SNPs with R-sq. <0.1. We show that this threshold is not appropriate and demonstrate that the magnitude of the problem is even greater with large sample size. We conclude that caution must be exercised when interpreting GWAS results derived from very large data sets showing strong evidence in support of epistatic interactions between markers.

genetics↗

bioRxiv2018

HaploBlocker: Creation of subgroup specific haplotype blocks and libraries

The concept of haplotype blocks has been shown to be useful in genetics. Fields of application range from the detection of regions under positive selection to statistical methods that make use of dimension reduction. We propose a novel approach (\"HaploBlocker\") for defining and inferring haplotype blocks that focuses on linkage instead of the commonly used population-wide measures of linkage disequilibrium (LD) which fail to identify segments shared by individuals in only a subset of the population. We define a haplotype block as a sequence of alleles that has a predefined minimum frequency in the population and only haplotypes with a similar sequence of alleles are considered to be carrying that block, effectively screening a dataset for group-wise identity-by-descent (IBD). Different to most other approaches these blocks are not restricted to shared start or end positions, but can overlap or even contain each other. From these haplotype blocks we construct a haplotype library that represents a large proportion of genetic variability of a population with a limited number of blocks. Our method is implemented in the associated R-package HaploBlocker and provides flexibility to not only optimize the structure of the obtained haplotype library for subsequent analyses (e.g., identification of shared segments between different populations), but is also able to handle datasets of different marker density and genetic diversity. By using haplotype blocks instead of SNPs, local epistatic interactions can be naturally modelled and the reduced number of parameter enables a wide variety of new methods for further genomic analyses. We illustrate our methodology with a dataset comprising 501 doubled haploid lines in a European maize landrace genotyped at 501124 SNPs. With the suggested approach, we identified 2851 haplotype blocks with an average length of 2633 SNPs (compared to 27.8 SNPs per block in HaploView) that together represent 94% of the dataset.\n\nAuthor summaryWhereas it is quite easy to identify segments of shared DNA between pairs of individuals, the problem becomes far more complex when analyzing a population. Especially for livestock and crop populations under strong selection one can observe long and possibly favourable segments that are segregating at high frequency. We propose here an adaptive and flexible approach to identify such segments (\"haplotype blocks\"). The main conceptual difference to other approaches is that we allow haplotype blocks to overlap so that patterns shared by a subset of the population can be mapped adequately. Afterwards, we select a set of those haplotype blocks that form a representation of the whole population (\"haplotype library\"). This haplotype library can be used similar to a SNP-dataset for subsequent genomic approaches with the advantage of a massive reduction of the number of parameters compared to standard haplotyping approaches. Since many breeding goals (e.g. grain yield, milk production) are known to be caused by complex interactions in genomic regions (or even the whole genome) using haplotype blocks instead of single base pairs provides a natural model for local interactions and enables the use of more complex models to incorporate distant interactions between genes, for instance.

genetics↗

bioRxiv2017

Accurate Genomic Prediction Of Human Height

We construct genomic predictors for heritable and extremely complex human quantitative traits (height, heel bone density, and educational attainment) using modern methods in high dimensional statistics (i.e., machine learning). Replication tests show that these predictors capture, respectively, ~40, 20, and 9 percent of total variance for the three traits. For example, predicted heights correlate ~0.65 with actual height; actual heights of most individuals in validation samples are within a few cm of the prediction. The variance captured for height is comparable to the estimated SNP heritability from GCTA (GREML) analysis, and seems to be close to its asymptotic value (i.e., as sample size goes to infinity), suggesting that we have captured most of the heritability for the SNPs used. Thus, our results resolve the common SNP portion of the \"missing heritability\" problem - i.e., the gap between prediction R-squared and SNP heritability. The ~20k activated SNPs in our height predictor reveal the genetic architecture of human height, at least for common SNPs. Our primary dataset is the UK Biobank cohort, comprised of almost 500k individual genotypes with multiple phenotypes. We also use other datasets and SNPs found in earlier GWAS for out-of-sample validation of our results.

genomics↗