Residual networks without pooling layers improve the accuracy of genomic predictions.

KEY MESSAGE: Residual neural network genomic selection is the first GS algorithm to reach 35 layers, and its prediction accuracy surpasses previous algorithms. With the decrease in DNA sequencing costs and the development of deep learning, phenotype prediction accuracy by genomic selection (GS) continues to improve. Residual networks, a widely validated deep learning technique, are introduced to deep learning for GS. Since each locus has a different weighted impact on the phenotype, strided convolutions are more suitable for GS problems than pooling layers. Through the above technological innovations, we propose a GS deep learning algorithm, residual neural network for genomic selection (ResGS). ResGS is the first neural network to reach 35 layers in GS. In 15 cases from four public data, the prediction accuracy of ResGS is higher than that of ridge-regression best linear unbiased prediction, support vector regression, random forest, gradient boosting regressor, and deep neural network genomic prediction in most cases. ResGS performs well in dealing with gene-environment interaction. Phenotypes from other environments are imported into ResGS along with genetic data. The prediction results are much better than just providing genetic data as input, which demonstrates the effectiveness of GS multi-modal learning. Standard deviation is recommended as an auxiliary GS evaluation metric, which could improve the distribution of predicted results. Deep learning for GS, such as ResGS, is becoming more accurate in phenotype prediction.

Clinical features, disease progression, and nuclear imaging in ATXN2-related parkinsonism in a longitudinal cohort.

BACKGROUND: Spinocerebellar ataxia 2 (SCA2) with a low range of CAG repeat expansion of ATXN2 gene can present with predominant or isolated parkinsonism that closely resembles Parkinson's disease (PD). This study is aimed at comparing clinical features, disease progression, and nuclear imaging between ATXN2-related parkinsonism (ATXN2-P) and PD. METHODS: Three hundred and seventy-seven clinically diagnosed PD with family history were screened by multiplex ligation-dependent probe amplification, whole-exome sequencing or target sequencing, and dynamic mutation testing of 10 SCA subtypes. The baseline and longitudinal clinical features as well as the dual-tracer positron emission tomography (PET) imaging were compared between ATXN2-P and genetically undefined familial PD (GU-fPD). RESULTS: Fifteen ATXN2-P patients from 7 families and 50 randomly selected GU-fPD patients were evaluated. Significantly less resting tremor and more symmetric signs were observed in ATXN2-P than GU-fPD. No significant difference was found in motor progression and duration from onset to occurrence of fluctuation, dyskinesia, and recurrent falls between the two groups. Cognitive impairment and rapid-eye-movement sleep behavior disorder were more common in ATXN2-P. During follow-up, olfaction was relatively spared, and no obvious progression of cognition dysfunction evaluated by Mini-Mental State Examination scores was found in ATXN2-P. PET results of ATXN2-P demonstrated a symmetric, diffuse, and homogenous dopamine transporter loss of bilateral striatum and a glucose metabolism pattern inconsistent with that in PD. CONCLUSIONS: Symmetric motor signs and unique nuclear imaging might be the clues to distinguish ATXN2-P from GU-fPD.

The genetic spectrum of a cohort of patients clinically diagnosed as Parkinson's disease in mainland China.

So far, over 20 causative genes of monogenic Parkinson's disease (PD) have been identified. Some causative genes of non-parkinsonian entities may also manifest with parkinsonism mimicking PD. This study aimed to investigate the genetic characteristics of clinically diagnosed PD with early onset age or family history. A total of 832 patients initially diagnosed with PD were enrolled, of which, 636 were classified into the early-onset group and 196 were classified into the familial late-onset group. The genetic testing included the multiplex ligation-dependent probe amplification and next generation sequencing (target sequencing or whole-exome sequencing). The dynamic variants of spinocerebellar ataxia were tested in probands with family history. In the early-onset group, 30.03% of patients (191/636) harbored pathogenic/likely pathogenic (P/LP) variants in known PD-related genes (CHCHD2, DJ-1, GBA (heterozygous), LRRK2, PINK1, PRKN, PLA2G6, SNCA and VPS35). Variants in PRKN were the most prevalent, accounting for 15.72% of the early-onset patients, followed by GBA (10.22%), and PLA2G6 (1.89%). And 2.52% (16/636) had P/LP variants in causative genes of other diseases (ATXN3, ATXN2, GCH1, TH, MAPT, GBA (homozygous)). In the familial late-onset group, 8.67% of patients (17/196) carried P/LP variants in known PD-related genes (GBA (heterozygous), HTRA2, SNCA) and 2.04% (4/196) had P/LP variants in other genes (ATXN2, PSEN1, DCTN1). Heterozygous GBA variants (7.14%) were the most common genetic cause found in familial late-onset patients. Genetic testing is of vital importance in differential diagnosis especially in early-onset and familial PD. Our findings may also provide some clues to the nomenclature of genetic movement disorders.

Reproducible metabolic topographies associated with multiple system atrophy: Network and regional analyses in Chinese and American patient cohorts.

PURPOSE: Multiple system atrophy (MSA) is an atypical parkinsonian syndrome and often difficult to discriminate clinically from progressive supranuclear palsy (PSP) and Parkinson's disease (PD) in early stages. Although a characteristic metabolic brain network has been reported for MSA, it is unknown whether this network can provide a clinically useful biomarker in different centers. This study was aimed to identify and cross-validate MSA-related brain network and assess its ability for differential diagnosis and clinical correlations in Chinese and American patient cohorts. METHODS: We included 18F-FDG PET scans retrospectively from 128 clinically diagnosed parkinsonian patients (34 MSA, 34 PSP and 60 PD) and 40 normal subjects in China and in the USA. Using PET images from 20 moderate-stage MSA patients of parkinsonian subtype and 20 normal subjects in both centers, we reproduced MSA-related pattern (MSAPRP) of spatial covariance and estimated its reliability. MSAPRP scores were evaluated in assessing differential diagnosis among moderate- and early-stage MSA, PSP or PD patients and clinical correlations with disease severity. Regional metabolic differences were detected using statistical parameter mapping analysis. MSA-related network and regional topographies of metabolic abnormality were cross-validated between the Chinese and American cohorts. RESULTS: We generated a highly reliable MSAPRP characterized by decreased loading in inferior frontal cortex, striatum and cerebellum, and increased loading in sensorimotor, parietal and occipital cortices. MSAPRP scores discriminated between normal, MSA, PSP and PD subjects and correlated with standardized ratings of clinical stages and motor symptoms in MSA. High similarities in MSAPRPs, network scores and corresponding maps of metabolic abnormality were observed between two different cohorts. CONCLUSION: We have demonstrated reproducible metabolic topographies associated with MSA at both network and regional levels in two independent patient cohorts. Moreover, MSAPRP scores are sensitive for evaluating disease discrimination and clinical correlates. This study supports differential diagnosis of MSA regardless of different patient populations, PET scanners and imaging protocols.