DeepHP: A New Gastric Mucosa Histopathology Dataset for Helicobacter pylori Infection Diagnosis.

Emerging deep learning-based applications in precision medicine include computational histopathological analysis. However, there is a lack of the required training image datasets to generate classification and detection models. This phenomenon occurs mainly due to human factors that make it difficult to obtain well-annotated data. The present study provides a curated public collection of histopathological images (DeepHP) and a convolutional neural network model for diagnosing gastritis. Images from gastric biopsy histopathological exams were used to investigate the performance of the proposed model in detecting gastric mucosa with Helicobacter pylori infection. The DeepHP database comprises 394,926 histopathological images, of which 111 K were labeled as Helicobacter pylori positive and 283 K were Helicobacter pylori negative. We investigated the classification performance of three Convolutional Neural Network architectures. The models were tested and validated with two distinct image sets of 15% (59K patches) chosen randomly. The VGG16 architecture showed the best results with an Area Under the Curve of 0.998%. The results showed that CNN could be used to classify histopathological images from gastric mucosa with marked precision. Our model evidenced high potential and application in the computational pathology field.

Pharmacogeomic implications of population diversity in Latin America: TPMT and NUDT15 polymorphisms and thiopurine dosing.

TPMT and NUDT15 polymorphisms are major determinants of tolerance to thiopurine drugs used in leukemias and nonmalignant immunologic disorders. We adopted an extreme discordant phenotype approach to explore the impact of Native American versus European ancestry on the distribution of TPMT and NUDT15 polymorphisms, and inferred metabolic phenotypes in the 1000 Genomes Ad Mixed American superpopulation. Significant differences were observed in the distribution of TPMT and NUDT15 haplotypes (star alleles) between individuals with predominant (>70%) European versus Native ancestry. The largest difference is related to NUDT15 rs116855232. Based on the combined TPMT/NUDT15 metabolic phenotypes, the Clinical Pharmacogenetics Implementation Consortium recommendations for thiopurine dose adjustment applies to 40.1% of individuals with major Native American ancestry, compared to 12.8% of individuals with predominantly European ancestry. These findings may be relevant to the adoption and interpretation of pharmacogenetic tests for thiopurine drugs across Latin America peoples with different European and Native-American ancestries.

Distribution and linkage disequilibrium of the enhancer SNP rs5758550 among Latin American populations: influence of continental ancestry.

OBJECTIVES: A single nucleotide polymorphism (SNP), rs5758550, in a critical enhancer region downstream of the CYP2D6 promoter was proposed to modulate CYP2D6 activity, depending on its linkage disequilibrium (LD) with the common CYP2D6 SNP, rs16947. We examined the influence of individual biogeographical ancestry on the frequency distribution of rs5758550 and its LD with rs16947 in Latin American populations. We then inferred the impact of rs5758550 on the predictive accuracy of CYP2D6 metabolizer status based on CYP2D6 haplotypes. METHODS: The study cohorts consisted of the Admixed American (AMR) superpopulation of the 1000 Genomes Project (n = 347) plus an admixed Brazilian (BR) cohort (N = 224). Individual proportions of Native, African and European ancestry estimated by ADMIXTURE analysis, were used to design four sub-cohorts, in which one of the three ancestral roots predominated largely (>6 fold) over the other two: AMR-NAT and AMR-EUR, comprised 80 AMR individuals each, with >70% Native or >70% European ancestry, BR-EUR and BR-AFR comprised Brazilians with >90% European (n = 80) or >70% African ancestry (n = 64), respectively. CYP2D6 haplotypes were inferred based on 10 commonly reported CYPD6 variants with or without addition of the enhancer rs5758550 SNP, pairwise LD was assessed by the R parameter, and activity scores were used to infer CYP2D6 metabolizer status. RESULTS: Minor allele frequency (MAF) of all CYP2D6 SNPs, except the rare (<0.02) rs5030656 and rs35742688, differed significantly across sub-cohorts, whereas no difference was observed for rs5758550. The R values for LD between rs5758550 and rs16947 ranged from 0.15 (BR-AFR) to 0.85 (AMR-NAT), with intermediate values in the predominantly European sub-cohorts (0.34-0.67). As a consequence, distribution of CYP2D6 haplotypes containing the rs16947 SNP plus rs5758550 wild-type (A) or variant (G) allele differed markedly across sub-cohorts. Comparison of the CYP2D6 activity scores assigned to the wild-type (CYP2D6*1) and the rs16947-containing haplotypes with or without inclusion of rs5758550, showed that knowledge of the rs5758550 genotype has negligible impact on predicted CYP2D6 phenotypes in AMR-EUR and AMR-NAT, but affects prediction in 10.7 and 21.6% of BR-EUR and BR-AFR individuals, respectively. CONCLUSION: Collectively, the present results reveal potential pharmacogenomic (PGx) implications of the population diversity in Latin America, affecting a major drug-metabolizing pathway. Thus, the influence of enhancer rs5758550 on assignment of CYP2D6 metabolic phenotypes varies markedly, according to the individual proportions of Native, European and African ancestry. This conclusion reinforces the notion that extrapolation of PGx data across the heterogeneous Latin American is risky, if not inappropriate.

NAToRA, a relatedness-pruning method to minimize the loss of dataset size in genetic and omics analyses.

Genetic and omics analyses frequently require independent observations, which is not guaranteed in real datasets. When relatedness cannot be accounted for, solutions involve removing related individuals (or observations) and, consequently, a reduction of available data. We developed a network-based relatedness-pruning method that minimizes dataset reduction while removing unwanted relationships in a dataset. It uses node degree centrality metric to identify highly connected nodes (or individuals) and implements heuristics that approximate the minimal reduction of a dataset to allow its application to complex datasets. When compared with two other popular population genetics methodologies (PLINK and KING), NAToRA shows the best combination of removing all relatives while keeping the largest possible number of individuals in all datasets tested and also, with similar effects on the allele frequency spectrum and Principal Component Analysis than PLINK and KING. NAToRA is freely available, both as a standalone tool that can be easily incorporated as part of a pipeline, and as a graphical web tool that allows visualization of the relatedness networks. NAToRA also accepts a variety of relationship metrics as input, which facilitates its use. We also release a genealogies simulator software used for different tests performed in this study.