Previous studies have identified differences in DNA methylation in autistic individuals compared to neurotypical individuals. Yet, it is unclear if this extends to autistic traits—subclinical manifestation of autism features in the general population. Here, we investigate the association between DNA methylation at birth (cord blood), and scores on the Social and Communication Disorders Checklist (SCDC), a measure of autistic traits, in 701 8-year-olds, by conducting a methylome-wide association study (MWAS). We did not identify significant CpGs associated with SCDC. The most significant CpG site was cg14379490, on chromosome 9 (MWAS beta = − 1.78 ± 0.35, p value = 5.34 × 10−7). Using methylation data for autism in peripheral tissues, we did not identify a significant concordance in effect direction of CpGs with p value < 10−4 in the SCDC MWAS (binomial sign test, p value > 0.5). In contrast, using methylation data for autism from post-mortem brain tissues, we identify a significant concordance in effect direction of CpGs with a p value < 10−4 in the SCDC MWAS (binomial sign test, p value = 0.004). Supporting this, we observe an enrichment for genes that are dysregulated in the post-mortem autism brain (one-sided Wilcoxon rank-sum test, p value = 6.22 × 10−5). Finally, integrating genome-wide association study (GWAS) data for autism (n = 46,350) with mQTL maps from cord-blood (n = 771), we demonstrate that mQTLs of CpGs associated with SCDC scores at p value thresholds of 0.01 and 0.005 are significantly shifted toward lower p values in the GWAS for autism (p < 5 × 10−3). We provide additional support for this using a GWAS of SCDC, and demonstrate a lack of enrichment in a GWAS of Alzheimer’s disease. Our results highlight the shared cross-tissue methylation architecture of autism and autistic traits, and demonstrate that mQTLs associated with differences in DNA methylation associated with childhood autistic traits are enriched for common genetic variants associated with autism and autistic traits.

Traditional genetic studies of single traits may be unable to detect the pleiotropic effects involved in complex diseases. To detect the correlation that exists between several phenotypes involved in the same biological process, we introduce an original methodology to analyze sets of correlated phenotypes involved in the coagulation cascade in genome-wide association studies. The methodology consists of a two-stage process. First, we define new phenotypic meta-variables (linear combinations of the original phenotypes), named metaphenotypes, by applying Independent Component Analysis for the multivariate analysis of correlated phenotypes (i.e. the levels of coagulation pathway–related proteins). The resulting metaphenotypes integrate the information regarding the underlying biological process (i.e. thrombus/clot formation). Secondly, we take advantage of a family based Genome Wide Association Study to identify genetic elements influencing these metaphenotypes and consequently thrombosis risk. Our study utilized data from the GAIT Project (Genetic Analysis of Idiopathic Thrombophilia). We obtained 15 metaphenotypes, which showed significant heritabilities, ranging from 0.2 to 0.7. These results indicate the importance of genetic factors in the variability of these traits. We found 4 metaphenotypes that showed significant associations with SNPs. The most relevant were those mapped in a region near the HRG, FETUB and KNG1 genes. Our results are provocative since they show that the KNG1 locus plays a central role as a genetic determinant of the entire coagulation pathway and thrombus/clot formation. Integrating data from multiple correlated measurements through metaphenotypes is a promising approach to elucidate the hidden genetic mechanisms underlying complex diseases.

Studies have reported persistent symptoms in patients hospitalized for COVID-19 up to 6 months post-discharge; however, sequalae beyond 6 months are unknown. This study aimed to investigate the clinical status of COVID-19 patients one year after hospital discharge and describe the factors related to poor outcomes. We conducted a single-center, prospective, cohort study of patients in Le Havre hospital (France) between 1 March 2020 and 11 May 2020. Baseline characteristics were collected from medical charts (including KATZ index and Clinical Frailty scale (CFS)), and a remote assessment was conducted 12 months after discharge. The main outcomes were the scores of the physical and mental components (PCS and MCS) of the Short-Form 36 (SF-36) and performance on the one-minute sit-to-stand test (STST1′). Scores <50% of the predicted values were considered as poor, and univariate and multivariate analyses were undertaken to investigate factors related to poor outcomes. Remote assessment was performed for 128 of the 157 (82%) eligible patients. Twenty-two patients were admitted to the intensive care unit (ICU), 45 to the intermediate care unit (IU), and 61 to the general ward (GW). Patients who spent time in ICU were more independent and younger. A large proportion of the sample had poor physical (30%) and mental health (27%) and a poor functional exercise capacity (33%) at the remote assessment. Higher levels of frailty at admission and hospital discharge were, respectively, associated with a higher risk of poor functional exercise capacity (StdOR 3.64 (95%CI 1.39–10.72); p = 0.01) and a higher risk of poor mental health (StdOR 2.81 (95%CI 1.17–7.45); p = 0.03). Long-term outcomes following hospitalization for COVID-19 infection may be negative for at least one year after discharge. Remote follow-up assessment could be highly beneficial for COVID-19 patients.

Accurate estimation of hip joint center (HJC) position is crucial during gait analysis. HJC is obtained with predictive or functional methods. But in the functional method, there is no consensus on where to place the skin markers and which combination to use. The objective of this study was to analyze how different combinations of skin markers affect the estimation of HJC position relative to predictive methods. Forty-one healthy volunteers were included in this study; thirteen markers were placed on the pelvis and hip of each subject’s lower limbs. Various marker combinations were used to determine the HJC position based on ten calibration movement trials, captured by a motion capture system. The estimated HJC position for each combination was evaluated by focusing on the range and standard deviation of the mean norm values of HJC and the mean X, Y, Z coordinates of HJC for each limb. The combinations that produced the best estimates incorporated the markers on the pelvis and on proximal and easily identifiable muscles, with results close to predictive methods. The combination that excluded the markers on the pelvis was not robust in estimating the HJC position.