Evaluation of enzyme activity predictions for variants of unknown significance in Arylsulfatase A.

Continued advances in variant effect prediction are necessary to demonstrate the ability of machine learning methods to accurately determine the clinical impact of variants of unknown significance (VUS). Towards this goal, the ARSA Critical Assessment of Genome Interpretation (CAGI) challenge was designed to characterize progress by utilizing 219 experimentally assayed missense VUS in the Arylsulfatase A (ARSA) gene to assess the performance of community-submitted predictions of variant functional effects. The challenge involved 15 teams, and evaluated additional predictions from established and recently released models. Notably, a model developed by participants of a genetics and coding bootcamp, trained with standard machine-learning tools in Python, demonstrated superior performance among submissions. Furthermore, the study observed that state-of-the-art deep learning methods provided small but statistically significant improvement in predictive performance compared to less elaborate techniques. These findings underscore the utility of variant effect prediction, and the potential for models trained with modest resources to accurately classify VUS in genetic and clinical research.

Choosing Variant Interpretation Tools for Clinical Applications: Context Matters.

Pathogenicity predictors are computational tools that classify genetic variants as benign or pathogenic; this is currently a major challenge in genomic medicine. With more than fifty such predictors available, selecting the most suitable tool for clinical applications like genetic screening, molecular diagnostics, and companion diagnostics has become increasingly challenging. To address this issue, we have developed a cost-based framework that naturally considers the various components of the problem. This framework encodes clinical scenarios using a minimal set of parameters and treats pathogenicity predictors as rejection classifiers, a common practice in clinical applications where low-confidence predictions are routinely rejected. We illustrate our approach in four examples where we compare different numbers of pathogenicity predictors for missense variants. Our results show that no single predictor is optimal for all clinical scenarios and that considering rejection yields a different perspective on classifiers.

Towards a New, Endophenotype-Based Strategy for Pathogenicity Prediction in BRCA1 and BRCA2: In Silico Modeling of the Outcome of HDR/SGE Assays for Missense Variants.

The present limitations in the pathogenicity prediction of BRCA1 and BRCA2 (BRCA1/2) missense variants constitute an important problem with negative consequences for the diagnosis of hereditary breast and ovarian cancer. However, it has been proposed that the use of endophenotype predictions, i.e., computational estimates of the outcomes of functional assays, can be a good option to address this bottleneck. The application of this idea to the BRCA1/2 variants in the CAGI 5-ENIGMA international challenge has shown promising results. Here, we developed this approach, exploring the predictive performances of the regression models applied to the BRCA1/2 variants for which the values of the homology-directed DNA repair and saturation genome editing assays are available. Our results first showed that we can generate endophenotype estimates using a few molecular-level properties. Second, we show that the accuracy of these estimates is enough to obtain pathogenicity predictions comparable to those of many standard tools. Third, endophenotype-based predictions are complementary to, but do not outperform, those of a Random Forest model trained using variant pathogenicity annotations instead of endophenotype values. In summary, our results confirmed the usefulness of the endophenotype approach for the pathogenicity prediction of the BRCA1/2 missense variants, suggesting different options for future improvements.

Polygenic analysis suggests the involvement of calcium signaling in executive function in schizophrenia patients.

Cognitive deficits are increasingly recognized as a core dimension rather than a consequence of schizophrenia (SCZ). The previous evidence supports the hypothesis of shared genetic factors between SCZ and cognitive ability. The objective of this study was to test whether and to what extent the variation of disease-relevant neurocognitive function in a sample of SCZ patients from the previous clinical interventional studies can be explained by SCZ polygenic risk scores (PRSs) or by hypothesis-driven and biomedical PRSs. The previous studies have described associations of the SNAP25 gene with cognition in SCZ. Likewise, the enrichment of several calcium signaling-related gene sets has been reported by genome-wide association studies (GWAS) in SCZ. Hypothesis-driven PRSs were calculated on the basis of the SNAP-25 interactome and also for genes regulated by phorbol myristate acetate (PMA), an activator of the signal transduction of protein kinase C (PKC) enzymes. In a cohort of 127 SCZ patients who had completed a comprehensive neurocognitive test battery as part of the previous antipsychotic intervention studies, we investigated the association between neurocognitive dimensions and PRSs. The PRS for SCZ and SNAP-25-associated genes could not explain the variance of neurocognition in this cohort. At a p value threshold of 0.05, the PRS for PMA was able to explain 2% of the variance in executive function (p = 0.05, uncorrected). The correlation between the PRS for PMA-regulated genes and cognition can give hints for further patient-derived cellular assays. In conclusion, incorporating biological information into PRSs and other en masse genetic analyses may help to close the gap between genetic vulnerability and the biological processes underlying neuropsychiatric diseases such as SCZ.

Assessment of blind predictions of the clinical significance of BRCA1 and BRCA2 variants.

Testing for variation in BRCA1 and BRCA2 (commonly referred to as BRCA1/2), has emerged as a standard clinical practice and is helping countless women better understand and manage their heritable risk of breast and ovarian cancer. Yet the increased rate of BRCA1/2 testing has led to an increasing number of Variants of Uncertain Significance (VUS), and the rate of VUS discovery currently outpaces the rate of clinical variant interpretation. Computational prediction is a key component of the variant interpretation pipeline. In the CAGI5 ENIGMA Challenge, six prediction teams submitted predictions on 326 newly-interpreted variants from the ENIGMA Consortium. By evaluating these predictions against the new interpretations, we have gained a number of insights on the state of the art of variant prediction and specific steps to further advance this state of the art.

BRCA1- and BRCA2-specific in silico tools for variant interpretation in the CAGI 5 ENIGMA challenge.

BRCA1 and BRCA2 (BRCA1/2) germline variants disrupting the DNA protective role of these genes increase the risk of hereditary breast and ovarian cancers. Correct identification of these variants then becomes clinically relevant, because it may increase the survival rates of the carriers. Unfortunately, we are still unable to systematically predict the impact of BRCA1/2 variants. In this article, we present a family of in silico predictors that address this problem, using a gene-specific approach. For each protein, we have developed two tools, aimed at predicting the impact of a variant at two different levels: Functional and clinical. Testing their performance in different datasets shows that specific information compensates the small number of predictive features and the reduced training sets employed to develop our models. When applied to the variants of the BRCA1/2 (ENIGMA) challenge in the fifth Critical Assessment of Genome Interpretation (CAGI 5) we find that these methods, particularly those predicting the functional impact of variants, have a good performance, identifying the large compositional bias towards neutral variants in the CAGI sample. This performance is further improved when incorporating to our prediction protocol estimates of the impact on splicing of the target variant.