1. Serra G, Mainas F, Golosio B, Retico A, Oliva P. Effect of data harmonization of multicentric dataset in ASD/TD classification. Brain informatics. 2023; 10(1): 32.

Machine Learning (ML) is nowadays an essential tool in the analysis of Magnetic Resonance Imaging (MRI) data, in particular in the identification of brain correlates in neurological and neurodevelopmental disorders. ML requires datasets of appropriate size for training, which in neuroimaging are typically obtained collecting data from multiple acquisition centers. However, analyzing large multicentric datasets can introduce bias due to differences between acquisition centers. ComBat harmonization is commonly used to address batch effects, but it can lead to data leakage when the entire dataset is used to estimate model parameters. In this study, structural and functional MRI data from the Autism Brain Imaging Data Exchange (ABIDE) collection were used to classify subjects with Autism Spectrum Disorders (ASD) compared to Typical Developing controls (TD). We compared the classical approach (external harmonization) in which harmonization is performed before train/test split, with an harmonization calculated only on the train set (internal harmonization), and with the dataset with no harmonization. The results showed that harmonization using the whole dataset achieved higher discrimination performance, while non-harmonized data and harmonization using only the train set showed similar results, for both structural and connectivity features. We also showed that the higher performances of the external harmonization are not due to larger size of the sample for the estimation of the model and hence these improved performance with the entire dataset may be ascribed to data leakage. In order to prevent this leakage, it is recommended to define the harmonization model solely using the train set.

Lien vers le texte intégral (Open Access ou abonnement)

2. Zhu J, Meng H, Li Y. Screening and Bioinformatics Analysis of Differential Genes in Autism Spectrum Disorder Based on GEO Database. Studies in health technology and informatics. 2023; 308: 280-8.

OBJECTIVE: The prevalence of autism spectrum disorder (ASD) in children has been increasing year by year, which has seriously affected the quality of life of children. There are many theories about the cause of ASDs, with some studies suggesting that it may be related to gene expression levels or inflammation and immune system dysfunction. But the exact mechanism is not fully understood. METHODS: profile of gene expression The protein interaction network (PPI) of differentially expressed genes was created using the STRING web tool and GSE77103, which was chosen from the gene expression omnibus (GEO) database. Using the CytoHubba plugin of Cytoscape program, the hub genes were examined. The hub gene regulatory network for miRNA-mRNA was then built. RESULTS: We identified 551 differentially expressed genes(DEGs) in 8 children with ASD and normal children. In addition, we screened out 10 hub genes (MX1, ISG15, IRF7, DDX58, IFIT1, BCL2L1, HPGDS, CTSD, PTGS2 and CD68) that were most associated with the development of ASDs. Then, microRNAs (miRNAs) closely related to hub genes (such as has-miR-27a-5p) were screened, and the miRNA-mRNA regulatory network was constructed. CONCLUSION: In this study, a total of 10 hub genes were identified, including MX1, ISG15, IRF7, DDX58, IFIT1, BCL2L1, HPGDS, CTSD, PTGS2 and CD68, which are closely related to ASD. These genes may play a key role in the occurrence and progression of ASD. In addition, we also revealed some miRNAs that regulate the hub genes of ASD. These results may deepen our understanding of ASD and provide potential biomarkers and targets for future treatment of patients with ASD.

Lien vers le texte intégral (Open Access ou abonnement)