Biology subjects

Krawitz, P.

Publications and source records attributed to Krawitz, P..

4 recordsLinked to original sources

bioRxiv2022

Boosting polygenic risk scores

Polygenic risk scores (PRS) evaluate the individual genetic liability to a certain trait and are expected to play an increasingly important role in the field of clinical risk stratification. Most often, PRS are estimated based on summary statistics of univariate effects derived from genome-wide association studies. To improve the predictive performance of PRS, it is desirable to fit multivariable models directly on the genetic data. Due to the large and high-dimensional data, a direct application of existing methods is often not feasible and new efficient algorithms are required to overcome the computational burden regarding efficiency and memory demands. We develop an adapted component-wise L2-boosting algorithm to fit genotype data from large cohort studies to continuous outcomes using linear base-learners for the genetic variants. Similar to the snpnet approach implementing lasso regression, the proposed snpboost approach iteratively works on smaller batches of variants. By restricting the set of possible base-learners in each boosting step to variants most correlated with the residuals from previous iterations, the computational efficiency can be substantially increased without losing prediction accuracy. Furthermore, for large-scale data based on various traits from the UK Biobank we show that our method yields competitive prediction accuracy and computational efficiency compared to the snpnet approach. Due to the modular structure of boosting, our framework can be further extended to construct PRS for different outcome data and effect types.

bioinformatics↗

bioRxiv2022

Statistical learning for sparser fine-mapped polygenic models: the prediction of LDL-cholesterol

Polygenic risk scores quantify the individual genetic predisposition regarding a particular trait. We propose and illustrate the application of existing statistical learning methods to derive sparser models for genome-wide data with a polygenic signal. Our approach is based on three consecutive steps. First, potentially informative loci are identified by a marginal screening approach. Then, fine-mapping is independently applied for blocks of variants in linkage disequilibrium, where informative variants are retrieved by using variable selection methods including boosting with probing and stochastic searches with the Adaptive Subspace method. Finally, joint prediction models with the selected variants are derived using statistical boosting. In contrast to alternative approaches relying on univariate summary statistics from genome-wide association studies, our three-step approach enables to select and fit multivariable regression models on large-scale genotype data. Based on UK Biobank data, we develop prediction models for LDL-cholesterol as a continuous trait. Additionally, we consider a recent scalable algorithm for the Lasso. Results show that statistical learning approaches based on fine-mapping of genetic signals result in a competitive prediction performance compared to classical polygenic risk approaches, while yielding sparser risk models that tend to be more robust regarding deviations from the target population.

genetics↗

bioRxiv2021

Drop of Prevalence after Population Expansion: A lower prevalence for recessive disorders in a random mating population is a transient phenomenon during and after a growth phase

Despite increasing data from population-wide sequencing studies, the risk for recessive disorders in consanguineous partnerships is still heavily debated. An important aspect that has not sufficiently been investigated theoretically, is the influence of inbreeding on mutation load and incidence rates when the population sizes change. We therefore developed a model to study these dynamics for a wide range of growth and mating conditions. In the phase of population expansion and shortly afterwards, our simulations show that there is a drop of diseased individuals at the expense of an increasing mutation load for random mating, while both parameters remain almost constant in highly consanguineous partnerships. This explains the empirical observation in present times that a high degree of consanguinity is associated with an increased risk of autosomal recessive disorders. However, it also states that the higher frequency of severe recessive disorders with developmental delay in inbred populations is a transient phenomenon before a mutation-selection balance is reached again. Author summaryWhat determines the recessive disease burden? The empirical observation that the proportion of intellectual disability of autosomal recessive cause is usually higher in offspring of consanguineous partnerships may lead to a misunderstanding about the mechanisms at work. In any population, selection removes pathogenic alleles from the gene pool while the de novo mutation rate adds novel pathogenic alleles. For comparable mutation rates, the incidence of severe recessive disease should be comparable in mutation-selection balance, regardless of the mating scheme. Different incidences can therefore only be explained by population dynamics that are not in equilibrium. We studied in simulated populations the time scales in which mutation-selection balance is reached after a growth phase and found that the mating scheme has a big impact on this lag time. When cousins mate preferentially with cousins, a few generations after a ten-fold increase in size the new equilibrium is established. In contrast, for random-mating, the transient advantage of a lower incidence may last for hundreds of generations, while the mutation load increases. By this means, our findings also highlight the importance of better carrier screens in the future for genetic consultations.

genetics↗

bioRxiv2018

PEDIA: Prioritization of Exome Data by Image Analysis

Phenotype information is crucial for the interpretation of genomic variants. So far it has only been accessible for bioinformatics workflows after encoding into clinical terms by expert dysmorphologists. Here, we introduce an approach, driven by artificial intelligence that uses portrait photographs for the interpretation of clinical exome data. We measured the value added by computer-assisted image analysis to the diagnostic yield on a cohort consisting of 679 individuals with 105 different monogenic disorders. For each case in the cohort we compiled frontal photos, clinical features and the disease-causing mutations and simulated multiple exomes of different ethnic backgrounds. With the additional use of similarity scores from computer-assisted analysis of frontal photos, we were able to achieve a top-10-accuracy rate for the disease-causing gene of 99 %. As this performance is significantly higher than without the information from facial pattern recognition, we make gestalt scores available for prioritization via an API.

bioinformatics↗