Predicting amplification of MYCN using CpG methylation biomarkers in neuroblastoma.

Background: Neuroblastoma is the most common extracranial solid tumor in childhood. Amplification of MYCN in neuroblastoma is a predictor of poor prognosis. Materials and methods: DNA methylation data from the TARGET data matrix were stratified into MYCN amplified and non-amplified groups. Differential methylation analysis, clustering, recursive feature elimination (RFE), machine learning (ML), Cox regression analysis and Kaplan-Meier estimates were performed. Results and Conclusion: 663 CpGs were differentially methylated between the two groups. A total of 25 CpGs were selected by RFE for clustering and ML, and a 100% clustering accuracy was obtained. ML validation on three external datasets produced high accuracy scores of 100%, 97% and 93%. Eight survival-associated CpGs were also identified. Therapeutic interventions may need to be targeted to patient subgroups.

Lay abstract Neuroblastoma is the most common extracranial solid tumor in childhood. Elevated levels of the MYCN protein in neuroblastoma is a predictor of poor prognosis. It is the most relevant prognostic factor in neuroblastoma and predicting MYCN gene amplification (which leads to increased gene expression and more protein) from epigenetic data rather than genetic testing might be useful in the oncology clinic. This study was designed to identify a DNA methylation (epigenetic) signature that can be used to diagnose MYCN amplification without actually testing for the gene. The authors also aimed to correlate this DNA methylation signature with patient survival and poorer prognosis. Based on statistical and computational methods applied to DNA methylation data for neuroblastoma, signatures that are predictive of MYCN amplification and poor prognosis were found, which clinicians can use for early patient diagnosis and selection of the best therapies for patients at high risk.

Developing reproducible bioinformatics analysis workflows for heterogeneous computing environments to support African genomics.

BACKGROUND: The Pan-African bioinformatics network, H3ABioNet, comprises 27 research institutions in 17 African countries. H3ABioNet is part of the Human Health and Heredity in Africa program (H3Africa), an African-led research consortium funded by the US National Institutes of Health and the UK Wellcome Trust, aimed at using genomics to study and improve the health of Africans. A key role of H3ABioNet is to support H3Africa projects by building bioinformatics infrastructure such as portable and reproducible bioinformatics workflows for use on heterogeneous African computing environments. Processing and analysis of genomic data is an example of a big data application requiring complex interdependent data analysis workflows. Such bioinformatics workflows take the primary and secondary input data through several computationally-intensive processing steps using different software packages, where some of the outputs form inputs for other steps. Implementing scalable, reproducible, portable and easy-to-use workflows is particularly challenging. RESULTS: H3ABioNet has built four workflows to support (1) the calling of variants from high-throughput sequencing data; (2) the analysis of microbial populations from 16S rDNA sequence data; (3) genotyping and genome-wide association studies; and (4) single nucleotide polymorphism imputation. A week-long hackathon was organized in August 2016 with participants from six African bioinformatics groups, and US and European collaborators. Two of the workflows are built using the Common Workflow Language framework (CWL) and two using Nextflow. All the workflows are containerized for improved portability and reproducibility using Docker, and are publicly available for use by members of the H3Africa consortium and the international research community. CONCLUSION: The H3ABioNet workflows have been implemented in view of offering ease of use for the end user and high levels of reproducibility and portability, all while following modern state of the art bioinformatics data processing protocols. The H3ABioNet workflows will service the H3Africa consortium projects and are currently in use. All four workflows are also publicly available for research scientists worldwide to use and adapt for their respective needs. The H3ABioNet workflows will help develop bioinformatics capacity and assist genomics research within Africa and serve to increase the scientific output of H3Africa and its Pan-African Bioinformatics Network.

Assessing computational genomics skills: Our experience in the H3ABioNet African bioinformatics network.

The H3ABioNet pan-African bioinformatics network, which is funded to support the Human Heredity and Health in Africa (H3Africa) program, has developed node-assessment exercises to gauge the ability of its participating research and service groups to analyze typical genome-wide datasets being generated by H3Africa research groups. We describe a framework for the assessment of computational genomics analysis skills, which includes standard operating procedures, training and test datasets, and a process for administering the exercise. We present the experiences of 3 research groups that have taken the exercise and the impact on their ability to manage complex projects. Finally, we discuss the reasons why many H3ABioNet nodes have declined so far to participate and potential strategies to encourage them to do so.

Assessment of the progression of kidney renal clear cell carcinoma using transcriptional profiles revealed new cancer subtypes with variable prognosis.

Background: Kidney renal clear cell carcinoma is the most prevalent subtype of renal cell carcinoma encompassing a heterogeneous group of malignancies. Accurate subtype identification and an understanding of the variables influencing prognosis are critical for personalized treatment, but currently limited. To facilitate the sub-classification of KIRC patients and improve prognosis, this study implemented a normalization method to track cancer progression by detecting the accumulation of genetic changes that occur throughout the multi-stage of cancer development. Objective: To reveal KIRC patients with different progression based on gene expression profiles using a normalization method. The aim is to refine molecular subtyping of KIRC patients associated with survival outcomes. Methods: RNA-sequenced gene expression of eighty-two KIRC patients were downloaded from UCSC Xena database. Advanced-stage samples were normalized with early-stage to account for differences in the multi-stage cancer progression's heterogeneity. Hierarchical clustering was performed to reveal clusters that progress differently. Two techniques were applied to screen for significant genes within the clusters. First, differentially expressed genes (DEGs) were discovered by Limma, thereafter, an optimal gene subset was selected using Recursive Feature Elimination (RFE). The gene subset was subjected to Random Forest Classifier to evaluate the cluster prediction performance. Genes strongly associated with survival were identified utilizing Cox regression analysis. The model's accuracy was assessed with Kaplan-Meier (K-M). Finally, a Gene ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analyses were performed. Results: Three clusters were revealed and categorized based on patients' overall survival into short, intermediate, and long. A total of 231 DEGs were discovered of which RFE selected 48 genes. Random Forest Classifier revealed a 100% cluster prediction performance of the genes. Five genes were identified with significant diagnostic capacity. The downregulation of genes SALL4 and KRT15 were associated with favorable prognosis, while the upregulation of genes OSBPL11, SPATA18, and TAL2 were associated with favorable prognosis. Conclusion: The normalization method based on tumour progression from early to late stages of cancer development revealed the heterogeneity of KIRC and identified three potential new subtypes with different prognoses. This could be of great importance for the development of new targeted therapies for each subtype.

Investigation of distinct gene expression profile patterns that can improve the classification of intermediate-risk prognosis in AML patients.

Background: Acute myeloid leukemia (AML) is a heterogeneous type of blood cancer that generally affects the elderly. AML patients are categorized with favorable-, intermediate-, and adverse-risks based on an individual's genomic features and chromosomal abnormalities. Despite the risk stratification, the progression and outcome of the disease remain highly variable. To facilitate and improve the risk stratification of AML patients, the study focused on gene expression profiling of AML patients within various risk categories. Therefore, the study aims to establish gene signatures that can predict the prognosis of AML patients and find correlations in gene expression profile patterns that are associated with risk groups. Methods: Microarray data were obtained from Gene Expression Omnibus (GSE6891). The patients were stratified into four subgroups based on risk and overall survival. Limma was applied to screen for differentially expressed genes (DEGs) between short survival (SS) and long survival (LS). DEGs strongly related to general survival were discovered using Cox regression and LASSO analysis. To assess the model's accuracy, Kaplan-Meier (K-M) and receiver operating characteristic (ROC) were used. A one-way ANOVA was performed to assess for differences in the mean gene expression profiles of the identified prognostic genes between the risk subcategories and survival. GO and KEGG enrichment analyses were performed on DEGs. Results: A total of 87 DEGs were identified between SS and LS groups. The Cox regression model selected nine genes CD109, CPNE3, DDIT4, INPP4B, LSP1, CPNE8, PLXNC1, SLC40A1, and SPINK2 that are associated with AML survival. K-M illustrated that the high expression of the nine-prognostic genes is associated with poor prognosis in AML. ROC further provided high diagnostic efficacy of the prognostic genes. ANOVA also validated the difference in gene expression profiles of the nine genes between the survival groups, and highlighted four prognostic genes to provide novel insight into risk subcategories poor and intermediate-poor, as well as good and intermediate-good that displayed similar expression patterns. Conclusion: Prognostic genes can provide more accurate risk stratification in AML. CD109, CPNE3, DDIT4, and INPP4B provided novel targets for better intermediate-risk stratification. This could enhance treatment strategies for this group, which constitutes the majority of adult AML patients.

Transforming RNA-Seq gene expression to track cancer progression in the multi-stage early to advanced-stage cancer development.

BACKGROUND: Cancer progression can be tracked by gene expression changes that occur throughout early-stage to advanced-stage cancer development. The accumulated genetic changes can be detected when gene expression levels in advanced-stage are less variable but show high variability in early-stage. Normalizing advanced-stage expression samples with early-stage and clustering of the normalized expression samples can reveal cancers with similar or different progression and provide insight into clinical and phenotypic patterns of patient samples within the same cancer. OBJECTIVE: This study aims to investigate cancer progression through RNA-Seq expression profiles across the multi-stage process of cancer development. METHODS: RNA-sequenced gene expression of Diffuse Large B-cell Lymphoma, Lung cancer, Liver cancer, Cervical cancer, and Testicular cancer were downloaded from the UCSC Xena database. Advanced-stage samples were normalized with early-stage samples to consider heterogeneity differences in the multi-stage cancer progression. WGCNA was used to build a gene network and categorized normalized genes into different modules. A gene set enrichment analysis selected key gene modules related to cancer. The diagnostic capacity of the modules was evaluated after hierarchical clustering. RESULTS: Unnormalized RNA-Seq gene expression failed to segregate advanced-stage samples based on selected cancer cohorts. Normalization with early-stage revealed the true heterogeneous gene expression that accumulates across the multi-stage cancer progression, this resulted in well segregated cancer samples. Cancer-specific pathways were enriched in the normalized WGCNA modules. The normalization method was further able to stratify patient samples based on phenotypic and clinical information. Additionally, the method allowed for patient survival analysis, with the Cox regression model selecting gene MAP4K1 in cervical cancer and Kaplan-Meier confirming that upregulation is favourable. CONCLUSION: The application of the normalization method further enhanced the accuracy of clustering of cancer samples based on how they progressed. Additionally, genes responsible for cancer progression were discovered.

Upregulation of FHL1, SPNS3, and MPZL2 predicts poor prognosis in pediatric acute myeloid leukemia patients with FLT3-ITD mutation.

Chromosomal translocations and gene mutations are characteristics of the genomic profile of acute myeloid leukemia (AML). We aim to identify a gene signature associated with poor prognosis in AML patients with FLT3-ITD compared to AML patients with NPM1/CEBPA mutations. RNA-sequencing (RNA-Seq) count data were downloaded from the UCSC Xena browser. Samples were grouped by their mutation status into high and low-risk groups. Differential gene expression (DGE), machine learning (ML) and survival analyses were performed. A total of 471 differentially expressed genes (DEGs) were identified, of which 16 DEGs were used as features for the prediction of mutation status. An accuracy of 92% was obtained from the ML model. FHL1, SPNS3, and MPZL2 were found to be associated with overall survival in FLT3-ITD samples. FLT3-ITD mutation confers an indicative gene expression profile different from NPM1/CEBPA mutation, and the expression of FHL1, SPSN3, and MPZL2 can serve as prognostic indicators of unfavorable disease.

Changes in subcutaneous adipose tissue microRNA expression in response to exercise training in obese African women.

The mechanisms that underlie exercise-induced adaptations in adipose tissue have not been elucidated, yet, accumulating studies suggest an important role for microRNAs (miRNAs). This study aimed to investigate miRNA expression in gluteal subcutaneous adipose tissue (GSAT) in response to a 12-week exercise intervention in South African women with obesity, and to assess depot-specific differences in miRNA expression in GSAT and abdominal subcutaneous adipose tissue (ASAT). In addition, the association between exercise-induced changes in miRNA expression and metabolic risk was evaluated. Women underwent 12-weeks of supervised aerobic and resistance training (n = 19) or maintained their regular physical activity during this period (n = 12). Exercise-induced miRNAs were identified in GSAT using Illumina sequencing, followed by analysis of differentially expressed miRNAs in GSAT and ASAT using quantitative real-time PCR. Associations between the changes (pre- and post-exercise training) in miRNA expression and metabolic parameters were evaluated using Spearman's correlation tests. Exercise training significantly increased the expression of miR-155-5p (1.5-fold, p = 0.045), miR-329-3p (2.1-fold, p < 0.001) and miR-377-3p (1.7-fold, p = 0.013) in GSAT, but not in ASAT. In addition, a novel miRNA, MYN0617, was identified in GSAT, with low expression in ASAT. The exercise-induced differences in miRNA expression were correlated with each other and associated with changes in high-density lipoprotein concentrations. Exercise training induced adipose-depot specific miRNA expression within subcutaneous adipose tissue depots from South African women with obesity. The significance of the association between exercise-induced miRNAs and metabolic risk warrants further investigation.

Effect of polyethylene glycol 20 000 on protein extraction efficiency of formalin-fixed paraffin-embedded tissues in South Africa.

BACKGROUND: Optimal protocols for efficient and reproducible protein extraction from formalin-fixed paraffin-embedded (FFPE) tissues are not yet standardised and new techniques are continually developed and improved. The effect of polyethylene glycol (PEG) 20 000 on protein extraction efficiency has not been evaluated using human FFPE colorectal cancer tissues and there is no consensus on the protein extraction solution required for efficient, reproducible extraction. OBJECTIVE: The impact of PEG 20 000 on protein extraction efficiency, reproducibility and protein selection bias was evaluated using FFPE colonic tissue via liquid chromatography tandem mass spectrometry analysis. METHODS: This study was conducted from August 2017 to July 2019 using human FFPE colorectal carcinoma tissues from the Anatomical Pathology department at Tygerberg Hospital in South Africa. Samples were analysed via label-free liquid chromatography tandem mass spectrometry to determine the impact of using PEG 20 000 in the protein extraction solution. Data were assessed regarding peptide and protein identifications, method efficiency, reproducibility, protein characteristics and organisation relating to gene ontology categories. RESULTS: Polyethylene glycol 20 000 exclusion increased peptides and proteins identifications and the method was more reproducible compared to the samples processed with PEG 20 000. However, no differences were observed with regard to protein selection bias. We found that higher protein concentrations (> 10 µg) compromised the function of PEG. CONCLUSION: This study indicates that protocols generating high protein yields from human FFPE tissues would benefit from the exclusion of PEG 20 000 in the protein extraction solution.

Evaluation of Protein Purification Techniques and Effects of Storage Duration on LC-MS/MS Analysis of Archived FFPE Human CRC Tissues.

To elucidate cancer pathogenesis and its mechanisms at the molecular level, the collecting and characterization of large individual patient tissue cohorts are required. Since most pathology institutes routinely preserve biopsy tissues by standardized methods of formalin fixation and paraffin embedment, these archived FFPE tissues are important collections of pathology material that include patient metadata, such as medical history and treatments. FFPE blocks can be stored under ambient conditions for decades, while retaining cellular morphology, due to modifications induced by formalin. However, the effect of long-term storage, at resource-limited institutions in developing countries, on extractable protein quantity/quality has not yet been investigated. In addition, the optimal sample preparation techniques required for accurate and reproducible results from label-free LC-MS/MS analysis across block ages remains unclear. This study investigated protein extraction efficiency of 1, 5, and 10-year old human colorectal carcinoma resection tissue and assessed three different gel-free protein purification methods for label-free LC-MS/MS analysis. A sample size of n = 17 patients per experimental group (with experiment power = 0.7 and α = 0.05, resulting in 70% confidence level) was selected. Data were evaluated in terms of protein concentration extracted, peptide/protein identifications, method reproducibility and efficiency, sample proteome integrity (due to storage time), as well as protein/peptide distribution according to biological processes, cellular components, and physicochemical properties. Data are available via ProteomeXchange with identifier PXD017198. The results indicate that the amount of protein extracted is significantly dependent on block age (p < 0.0001), with older blocks yielding less protein than newer blocks. Detergent removal plates were the most efficient and overall reproducible protein purification method with regard to number of peptide and protein identifications, followed by the MagReSyn® SP3/HILIC method (with on-bead enzymatic digestion), and lastly the acetone precipitation and formic acid resolubilization method. Overall, the results indicate that long-term storage of FFPE tissues (as measured by methionine oxidation) does not considerably interfere with retrospective proteomic analysis (p > 0.1). Block age mainly affects initial protein extraction yields and does not extensively impact on subsequent label-free LC-MS/MS analysis results.

Identification of novel prognostic markers of survival time in high-risk neuroblastoma using gene expression profiles.

Neuroblastoma is the most common extracranial solid tumor in childhood. Patients in high-risk group often have poor outcomes with low survival rates despite several treatment options. This study aimed to identify a genetic signature from gene expression profiles that can serve as prognostic indicators of survival time in patients of high-risk neuroblastoma, and that could be potential therapeutic targets. RNA-seq count data was downloaded from UCSC Xena browser and samples grouped into Short Survival (SS) and Long Survival (LS) groups. Differential gene expression (DGE) analysis, enrichment analyses, regulatory network analysis and machine learning (ML) prediction of survival group were performed. Forty differentially expressed genes (DEGs) were identified including genes involved in molecular function activities essential for tumor proliferation. DEGs used as features for prediction of survival groups included EVX2, NHLH2, PRSS12, POU6F2, HOXD10, MAPK15, RTL1, LGR5, CYP17A1, OR10AB1P, MYH14, LRRTM3, GRIN3A, HS3ST5, CRYAB and NXPH3. An accuracy score of 82% was obtained by the ML classification models. SMIM28 was revealed to possibly have a role in tumor proliferation and aggressiveness. Our results indicate that these DEGs can serve as prognostic indicators of survival in high-risk neuroblastoma patients and will assist clinicians in making better therapeutic and patient management decisions.

Sperm Proteome Analysis and Identification of Fertility-Associated Biomarkers in Unexplained Male Infertility.

Up to 30% of men with normal semen parameters suffer from infertility and the reason for this is unknown. Altered expression of sperm proteins may be a major cause of infertility in these men. Proteomic profiling was performed on pooled semen samples from eight normozoospermic fertile men and nine normozoospermic infertile men using LC-MS/MS. Furthermore, key differentially expressed proteins (DEPs) related to the fertilization process were selected for validation using Western blotting. A total of 1139 and 1095 proteins were identified in normozoospermic fertile and infertile men, respectively. Of these, 162 proteins were identified as DEPs. The canonical pathway related to free radical scavenging was enriched with upregulated DEPs in normozoospermic infertile men. The proteins associated with reproductive system development and function, and the ubiquitination pathway were underexpressed in normozoospermic infertile men. Western blot analysis revealed the overexpression of annexin A2 (ANXA2) (2.03 fold change; P = 0.0243), and underexpression of sperm surface protein Sp17 (SPA17) (0.37 fold change; P = 0.0205) and serine protease inhibitor (SERPINA5) (0.32 fold change; P = 0.0073) in men with unexplained male infertility (UMI). The global proteomic profile of normozoospermic infertile men is different from that of normozoospermic fertile men. Our data suggests that SPA17, ANXA2, and SERPINA5 may potentially serve as non-invasive protein biomarkers associated with the fertilization process of the spermatozoa in UMI.

Organizing and running bioinformatics hackathons within Africa: The H3ABioNet cloud computing experience.

The need for portable and reproducible genomics analysis pipelines is growing globally as well as in Africa, especially with the growth of collaborative projects like the Human Health and Heredity in Africa Consortium (H3Africa). The Pan-African H3Africa Bioinformatics Network (H3ABioNet) recognized the need for portable, reproducible pipelines adapted to heterogeneous compute environments, and for the nurturing of technical expertise in workflow languages and containerization technologies. To address this need, in 2016 H3ABioNet arranged its first Cloud Computing and Reproducible Workflows Hackathon, with the purpose of building key genomics analysis pipelines able to run on heterogeneous computing environments and meeting the needs of H3Africa research projects. This paper describes the preparations for this hackathon and reflects upon the lessons learned about its impact on building the technical and scientific expertise of African researchers. The workflows developed were made publicly available in GitHub repositories and deposited as container images on quay.io.

Baobab Laboratory Information Management System: Development of an Open-Source Laboratory Information Management System for Biobanking.

A laboratory information management system (LIMS) is central to the informatics infrastructure that underlies biobanking activities. To date, a wide range of commercial and open-source LIMSs are available and the decision to opt for one LIMS over another is often influenced by the needs of the biobank clients and researchers, as well as available financial resources. The Baobab LIMS was developed by customizing the Bika LIMS software ( www.bikalims.org ) to meet the requirements of biobanking best practices. The need to implement biobank standard operation procedures as well as stimulate the use of standards for biobank data representation motivated the implementation of Baobab LIMS, an open-source LIMS for Biobanking. Baobab LIMS comprises modules for biospecimen kit assembly, shipping of biospecimen kits, storage management, analysis requests, reporting, and invoicing. The Baobab LIMS is based on the Plone web-content management framework. All the system requirements for Plone are applicable to Baobab LIMS, including the need for a server with at least 8 GB RAM and 120 GB hard disk space. Baobab LIMS is a server-client-based system, whereby the end user is able to access the system securely through the internet on a standard web browser, thereby eliminating the need for standalone installations on all machines.

NetCapDB: measuring bioinformatics capacity development in Africa.

BACKGROUND: The National Institutes of Health (USA) has committed 5 years of funding to the Bioinformatics Network of the Human Heredity and Health in Africa initiative. This pan-African network aims to develop capacity for bioinformatics research, in order to provide support to human health genomics research programs ongoing on the continent. Over the 5 years of funding, it is imperative to track changes in bioinformatics capacity at the funded centres and to document how the funding has translated into capacity development during this time frame. RESULTS: The Network capacity database, NetCapDB, is a relational database that captures quantitative metrics for bioinformatics capacity, and tracks the changes in these metrics over time. A graphical user interface allows for straight-forward, browser-based data entry by users across Africa; and for visual and graph-based exploration of captured data. A reporting interface allows for semi-automated generation of standardized reports for monitoring and evaluation purposes.