Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

General General

Identification of a deep intronic POLR3A variant causing inclusion of a pseudoexon derived from an Alu element in Pol III-related leukodystrophy.

In Journal of human genetics

Pseudoexon inclusion caused by deep intronic variants is an important genetic cause for various disorders. Here, we present a case of a hypomyelinating leukodystrophy with developmental delay, intellectual disability, autism spectrum disorder, and hypodontia, which are consistent with autosomal recessive POLR3-related leukodystrophy. Whole-exome sequencing identified only a heterozygous missense variant (c.1451G>A) in POLR3A. To explore possible involvement of a deep intronic variant in another allele, we performed whole-genome sequencing of the patient with variant annotation by SpliceAI, a deep-learning-based splicing prediction tool. A deep intronic variant (c.645 + 312C>T) in POLR3A, which was predicted to cause inclusion of a pseudoexon derived from an Alu element, was identified and confirmed by mRNA analysis. These results clearly showed that whole-genome sequencing, in combination with deep-learning-based annotation tools such as SpliceAI, will bring us further benefits in detecting and evaluating possible pathogenic variants in deep intronic regions.

Hiraide Takuya, Nakashima Mitsuko, Ikeda Takahiro, Tanaka Daisuke, Osaka Hitoshi, Saitsu Hirotomo

2020-Jun-01

General General

Effect of congenital adrenal hyperplasia treated by glucocorticoids on plasma metabolome: a machine-learning-based analysis.

In Scientific reports ; h5-index 158.0

BACKGROUND : Congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency leads to impaired cortisol biosynthesis. Treatment includes glucocorticoid supplementation. We studied the specific metabolomics signatures in CAH patients using two different algorithms.

METHODS : In a case-control study of CAH patients matched on sex and age with healthy control subjects, two metabolomic analyses were performed: one using MetaboDiff, a validated differential metabolomic analysis tool and the other, using Predomics, a novel machine-learning algorithm.

RESULTS : 168 participants were included (84 CAH patients). There was no correlation between plasma cortisol levels during glucocorticoid supplementation and metabolites in CAH patients. Indoleamine 2,3-dioxygenase enzyme activity was correlated with ACTH (rho coefficient = -0.25, p-value = 0.02), in CAH patients but not in controls subjects. Overall, 33 metabolites were significantly altered in CAH patients. Main changes came from: purine and pyrimidine metabolites, branched aminoacids, tricarboxylic acid cycle metabolites and associated pathways (urea, glucose, pentose phosphates). MetaboDiff identified 2 modules that were significantly different between both groups: aminosugar metabolism and purine metabolism. Predomics found several interpretable models which accurately discriminated the two groups (accuracy of 0.86 and AUROC of 0.9).

CONCLUSION : CAH patients and healthy control subjects exhibit significant differences in plasma metabolomes, which may be explained by glucocorticoid supplementation.

Nguyen Lee S, Prifti Edi, Ichou Farid, Leban Monique, Funck-Brentano Christian, Touraine Philippe, Salem Joe-Elie, Bachelot Anne

2020-Jun-01

General General

Machine learning uncovers cell identity regulator by histone code.

In Nature communications ; h5-index 260.0

Conversion between cell types, e.g., by induced expression of master transcription factors, holds great promise for cellular therapy. Our ability to manipulate cell identity is constrained by incomplete information on cell identity genes (CIGs) and their expression regulation. Here, we develop CEFCIG, an artificial intelligent framework to uncover CIGs and further define their master regulators. On the basis of machine learning, CEFCIG reveals unique histone codes for transcriptional regulation of reported CIGs, and utilizes these codes to predict CIGs and their master regulators with high accuracy. Applying CEFCIG to 1,005 epigenetic profiles, our analysis uncovers the landscape of regulation network for identity genes in individual cell or tissue types. Together, this work provides insights into cell identity regulation, and delivers a powerful technique to facilitate regenerative medicine.

Xia Bo, Zhao Dongyu, Wang Guangyu, Zhang Min, Lv Jie, Tomoiaga Alin S, Li Yanqiang, Wang Xin, Meng Shu, Cooke John P, Cao Qi, Zhang Lili, Chen Kaifu

2020-Jun-01

General General

Integrated omics in Drosophila uncover a circadian kinome.

In Nature communications ; h5-index 260.0

Most organisms on the earth exhibit circadian rhythms in behavior and physiology, which are driven by endogenous clocks. Phosphorylation plays a central role in timing the clock, but how this contributes to overt rhythms is unclear. Here we conduct phosphoproteomics in conjunction with transcriptomic and proteomic profiling using fly heads. By developing a pipeline for integrating multi-omics data, we identify 789 (~17%) phosphorylation sites with circadian oscillations. We predict 27 potential circadian kinases to participate in phosphorylating these sites, including 7 previously known to function in the clock. We screen the remaining 20 kinases for effects on circadian rhythms and find an additional 3 to be involved in regulating locomotor rhythm. We re-construct a signal web that includes the 10 circadian kinases and identify GASKET as a potentially important regulator. Taken together, we uncover a circadian kinome that potentially shapes the temporal pattern of the entire circadian molecular landscapes.

Wang Chenwei, Shui Ke, Ma Shanshan, Lin Shaofeng, Zhang Ying, Wen Bo, Deng Wankun, Xu Haodong, Hu Hui, Guo Anyuan, Xue Yu, Zhang Luoying

2020-Jun-01

Radiology Radiology

Fully automated segmentation of the right ventricle in patients with repaired Tetralogy of Fallot using U-Net.

In Proceedings of SPIE--the International Society for Optical Engineering

Cardiac magnetic resonance (CMR) imaging is considered the standard imaging modality for volumetric analysis of the right ventricle (RV), an especially important practice in the evaluation of heart structure and function in patients with repaired Tetralogy of Fallot (rTOF). In clinical practice, however, this requires time-consuming manual delineation of the RV endocardium in multiple 2-dimensional (2D) slices at multiple phases of the cardiac cycle. In this work, we employed a U-Net based 2D convolutional neural network (CNN) classifier in the fully automatic segmentation of the RV blood pool. Our dataset was comprised of 5,729 short-axis cine CMR slices taken from 100 individuals with rTOF. Training of our CNN model was performed on images from 50 individuals while validation was performed on images from 10 individuals. Segmentation results were evaluated by Dice similarity coefficient (DSC) and Hausdorff distance (HD). Use of the CNN model on our testing group of 40 individuals yielded a median DSC of 90% and a median 95th percentile HD of 5.1 mm, demonstrating good performance in these metrics when compared to literature results. Our preliminary results suggest that our deep learning-based method can be effective in automating RV segmentation.

Tran Christopher T, Halicek Martin, Dormer James D, Tandon Animesh, Hussain Tarique, Fei Baowei

2020-Feb

Cardiac magnetic resonance imaging, Convolutional neural network (CNN), Deep learning, Heart, Image segmentation, Left ventricle, Tetralogy of Fallot

Radiology Radiology

Segmentation of uterus and placenta in MR images using a fully convolutional neural network.

In Proceedings of SPIE--the International Society for Optical Engineering

Segmentation of the uterine cavity and placenta in fetal magnetic resonance (MR) imaging is useful for the detection of abnormalities that affect maternal and fetal health. In this study, we used a fully convolutional neural network for 3D segmentation of the uterine cavity and placenta while a minimal operator interaction was incorporated for training and testing the network. The user interaction guided the network to localize the placenta more accurately. We trained the network with 70 training and 10 validation MRI cases and evaluated the algorithm segmentation performance using 20 cases. The average Dice similarity coefficient was 92% and 82% for the uterine cavity and placenta, respectively. The algorithm could estimate the volume of the uterine cavity and placenta with average errors of 2% and 9%, respectively. The results demonstrate that the deep learning-based segmentation and volume estimation is possible and can potentially be useful for clinical applications of human placental imaging.

Shahedi Maysam, Dormer James D, T T Anusha Devi, Do Quyen N, Xi Yin, Lewis Matthew A, Madhuranthakam Ananth J, Twickler Diane M, Fei Baowei

2020-Feb

Convolutional neural network, fetal magnetic resonance imaging, image segmentation, placenta, uterus