Increasing metadata coverage of SRA BioSample entries using deep learning-based named entity recognition

High-quality metadata annotations for data hosted in large public repositories are essential for research reproducibility and for conducting fast, powerful and scalable meta-analyses. Currently, a majority of sequencing samples in the National Center for Biotechnology Information's Sequence Read Archive (SRA) are missing metadata across several categories. In an effort to improve the metadata coverage of these samples, we leveraged almost 44 million attribute-value pairs from SRA BioSample to train a scalable, recurrent neural network that predicts missing metadata via named entity recognition (NER). The network was first trained to classify short text phrases according to 11 metadata categories and achieved an overall accuracy and area under the receiver operating characteristic curve of 85.2% and 0.977, respectively. We then applied our classifier to predict 11 metadata categories from the longer TITLE attribute of samples, evaluating performance on a set of samples withheld from model training. Prediction accuracies were high when extracting sample Genus/Species (94.85%), Condition/Disease (95.65%) and Strain (82.03%) from TITLEs, with lower accuracies and lack of predictions for other categories highlighting multiple issues with the current metadata annotations in BioSample. These results indicate the utility of recurrent neural networks for NER-based metadata prediction and the potential for models such as the one presented here to increase metadata coverage in BioSample while minimizing the need for manual curation. Database URL: https://github.com/cartercompbio/PredictMEE.

Activation of NF-κB and p300/CBP potentiates cancer chemoimmunotherapy through induction of MHC-I antigen presentation

Many cancers evade immune rejection by suppressing major histocompatibility class I (MHC-I) antigen processing and presentation (AgPP). Such cancers do not respond to immune checkpoint inhibitor therapies (ICIT) such as PD-1/PD-L1 [PD-(L)1] blockade. Certain chemotherapeutic drugs augment tumor control by PD-(L)1 inhibitors through potentiation of T-cell priming but whether and how chemotherapy enhances MHC-I-dependent cancer cell recognition by cytotoxic T cells (CTLs) is not entirely clear. We now show that the lysine acetyl transferases p300/CREB binding protein (CBP) control MHC-I AgPPM expression and neoantigen amounts in human cancers. Moreover, we found that two distinct DNA damaging drugs, the platinoid oxaliplatin and the topoisomerase inhibitor mitoxantrone, strongly up-regulate MHC-I AgPP in a manner dependent on activation of nuclear factor kappa B (NF-κB), p300/CBP, and other transcription factors, but independently of autocrine IFNγ signaling. Accordingly, NF-κB and p300 ablations prevent chemotherapy-induced MHC-I AgPP and abrogate rejection of low MHC-I-expressing tumors by reinvigorated CD8+ CTLs. Drugs like oxaliplatin and mitoxantrone may be used to overcome resistance to PD-(L)1 inhibitors in tumors that had "epigenetically down-regulated," but had not permanently lost MHC-I AgPP activity.

Psychosocial difficulties, obesity and disadvantage in a cohort of Irish children

Background

Childhood obesity is stabilising in Ireland, but less so in children from disadvantaged backgrounds. There is also evidence for a relationship between psychosocial difficulties and obesity in youth, but findings are often mixed. This study aimed to describe the relationships between psychosocial difficulties, obesity, and disadvantage in a national cohort of children in Ireland.

Methods

Participants (n = 7275) were surveyed and measured at age nine and thirteen as part of the Growing Up in Ireland study (GUI). Psychosocial difficulties were measured using the Strengths and Difficulties Questionnaire (SDQ). Disadvantage was represented by the education status of the primary caregiver. BMI was calculated using measured height and weight. Weight status categories were created using the UK 1990 Growth Data. Correlation analyses were used to examine the relationship between child BMI and SDQ scores. Associations between weight categories, psychosocial difficulties and education status were examined using Chi-square analysis.

Results

In children whose parents had secondary education, the obesity rate was 27.9%, compared to 24.7% in those with a college degree (p < 0.01). This relationship was also significant at age thirteen (<0.01). Furthermore, BMI was positively correlated with SDQ total score, (r=.107, p = 0.01) at age nine and at age thirteen (r=.089, p = 0.01). A significant association was also found between weight status categories and psychosocial difficulties at ages nine (p < 0.1), and thirteen (p < 0.1). Further findings will be reported following a preliminary longitudinal analysis.

Conclusions

Obesity is associated with psychosocial difficulties, as well as the education status of the primary caregiver, in a large sample of Irish children. These findings are important for understanding the relationship between childhood obesity and psychological health and have implications for the treatment of obesity in young people.

Key messages

There are higher rates of obesity in children whose parents do not have a college degree. Childhood obesity remains a serious public health issue, and is associated with psychosocial difficulties in an Irish cohort of children.

Food-Intake Normalization of Dysregulated Fatty Acids in Women with Anorexia Nervosa

Anorexia nervosa (AN) is a psychiatric disorder affected by psychological, environmental, and biological factors. Individuals with AN avoid high-fat, high-calorie diets and have shown abnormal metabolism of fatty acids (FAs), which are essential for brain and cognitive/neuropsychiatric health. To clarify the relationship between FAs and AN, fasting and postprandial plasma FAs in AN patients and age-matched control women were analyzed via mass-spectrometry. Clinical phenotypes were assessed using Becker Anxiety Inventory and Becker Depression Inventory. AN patients and controls exhibited different FA signatures at both fasting and postprandial timepoints. Lauric acid, eicosapentaenoic acid (EPA), docosapentaenoic acid (DPA), and alpha-linoleic acid (ALA) were higher in AN than in controls (lauric acid: 15,081.6 ± 14,970.2 vs. 8257.4 ± 4740.2 pmol/mL; ALA at fasting: 2217.7 ± 1587.6 vs. 1087.9 ± 821.2 pmol/mL; ALA at postprandial: 1830.9 ± 1115.6 vs. 1159.4 ± 664.7 pmol/mL. EPA: 33,788.3 ± 17,487.5 vs. 22,860.6 ± 12,642.4 pmol/mL; DPA: 32,664.8 ± 16,215.0 vs. 20,969.0 ± 12,350.0 pmol/mL. FDR-adjusted p-values < 0.05). Food intake and AN status modified the correlations of FAs with body mass index (BMI), depression, and anxiety. Desaturases SCD-18 and D6D showed lower activities in AN compared to controls. Altered FA signature, specifically correlations between elevated n-3 FAs and worsened symptoms, illustrate metabolic underpinnings in AN. Future studies should investigate the mechanisms by which FA dysregulation, specifically elevated n-3 FAs, affects AN risk and outcome.

Integrative genomic analysis of mouse and human hepatocellular carcinoma

Cancer genomics has enabled the exhaustive molecular characterization of tumors and exposed hepatocellular carcinoma (HCC) as among the most complex cancers. This complexity is paralleled by dozens of mouse models that generate histologically similar tumors but have not been systematically validated at the molecular level. Accurate models of the molecular pathogenesis of HCC are essential for biomedical progress; therefore we compared genomic and transcriptomic profiles of four separate mouse models [MUP transgenic, TAK1-knockout, carcinogen-driven diethylnitrosamine (DEN), and Stelic Animal Model (STAM)] with those of 987 HCC patients with distinct etiologies. These four models differed substantially in their mutational load, mutational signatures, affected genes and pathways, and transcriptomes. STAM tumors were most molecularly similar to human HCC, with frequent mutations in Ctnnb1, similar pathway alterations, and high transcriptomic similarity to high-grade, proliferative human tumors with poor prognosis. In contrast, TAK1 tumors better reflected the mutational signature of human HCC and were transcriptionally similar to low-grade human tumors. DEN tumors were least similar to human disease and almost universally carried the Braf V637E mutation, which is rarely found in human HCC. Immune analysis revealed that strain-specific MHC-I genotype can influence the molecular makeup of murine tumors. Thus, different mouse models of HCC recapitulate distinct aspects of HCC biology, and their use should be adapted to specific questions based on the molecular features provided here.

The Emerging Potential for Network Analysis to Inform Precision Cancer Medicine

Precision cancer medicine promises to tailor clinical decisions to patients using genomic information. Indeed, successes of drugs targeting genetic alterations in tumors, such as imatinib that targets BCR-ABL in chronic myelogenous leukemia, have demonstrated the power of this approach. However, biological systems are complex, and patients may differ not only by the specific genetic alterations in their tumor, but also by more subtle interactions among such alterations. Systems biology and more specifically, network analysis, provides a framework for advancing precision medicine beyond clinical actionability of individual mutations. Here we discuss applications of network analysis to study tumor biology, early methods for N-of-1 tumor genome analysis, and the path for such tools to the clinic.

An evaluation of copy number variation detection tools for cancer using whole exome sequencing data

BACKGROUND

Recently copy number variation (CNV) has gained considerable interest as a type of genomic/genetic variation that plays an important role in disease susceptibility. Advances in sequencing technology have created an opportunity for detecting CNVs more accurately. Recently whole exome sequencing (WES) has become primary strategy for sequencing patient samples and study their genomics aberrations. However, compared to whole genome sequencing, WES introduces more biases and noise that make CNV detection very challenging. Additionally, tumors' complexity makes the detection of cancer specific CNVs even more difficult. Although many CNV detection tools have been developed since introducing NGS data, there are few tools for somatic CNV detection for WES data in cancer.

RESULTS

In this study, we evaluated the performance of the most recent and commonly used CNV detection tools for WES data in cancer to address their limitations and provide guidelines for developing new ones. We focused on the tools that have been designed or have the ability to detect cancer somatic aberrations. We compared the performance of the tools in terms of sensitivity and false discovery rate (FDR) using real data and simulated data. Comparative analysis of the results of the tools showed that there is a low consensus among the tools in calling CNVs. Using real data, tools show moderate sensitivity (~50% - ~80%), fair specificity (~70% - ~94%) and poor FDRs (~27% - ~60%). Also, using simulated data we observed that increasing the coverage more than 10× in exonic regions does not improve the detection power of the tools significantly.

CONCLUSIONS

The limited performance of the current CNV detection tools for WES data in cancer indicates the need for developing more efficient and precise CNV detection methods. Due to the complexity of tumors and high level of noise and biases in WES data, employing advanced novel segmentation, normalization and de-noising techniques that are designed specifically for cancer data is necessary. Also, CNV detection development suffers from the lack of a gold standard for performance evaluation. Finally, developing tools with user-friendly user interfaces and visualization features can enhance CNV studies for a broader range of users.

PREMIX: PRivacy-preserving EstiMation of Individual admiXture

In this paper we proposed a framework: PRivacy-preserving EstiMation of Individual admiXture (PREMIX) using Intel software guard extensions (SGX). SGX is a suite of software and hardware architectures to enable efficient and secure computation over confidential data. PREMIX enables multiple sites to securely collaborate on estimating individual admixture within a secure enclave inside Intel SGX. We implemented a feature selection module to identify most discriminative Single Nucleotide Polymorphism (SNP) based on informativeness and an Expectation Maximization (EM)-based Maximum Likelihood estimator to identify the individual admixture. Experimental results based on both simulation and 1000 genome data demonstrated the efficiency and accuracy of the proposed framework. PREMIX ensures a high level of security as all operations on sensitive genomic data are conducted within a secure enclave using SGX.

Genome privacy: challenges, technical approaches to mitigate risk, and ethical considerations in the United States

Accessing and integrating human genomic data with phenotypes are important for biomedical research. Making genomic data accessible for research purposes, however, must be handled carefully to avoid leakage of sensitive individual information to unauthorized parties and improper use of data. In this article, we focus on data sharing within the scope of data accessibility for research. Current common practices to gain biomedical data access are strictly rule based, without a clear and quantitative measurement of the risk of privacy breaches. In addition, several types of studies require privacy-preserving linkage of genotype and phenotype information across different locations (e.g., genotypes stored in a sequencing facility and phenotypes stored in an electronic health record) to accelerate discoveries. The computer science community has developed a spectrum of techniques for data privacy and confidentiality protection, many of which have yet to be tested on real-world problems. In this article, we discuss clinical, technical, and ethical aspects of genome data privacy and confidentiality in the United States, as well as potential solutions for privacy-preserving genotype-phenotype linkage in biomedical research.

Impact of reducing the threshold for acetylcysteine treatment in acute paracetamol poisoning: the recent United Kingdom experience

BACKGROUND

On 3 September 2012, the licensed indication for acetylcysteine was changed in the United Kingdom (UK) so that all patients with a plasma paracetamol concentration above a "100 mg/L" (4 h post ingestion) nomogram treatment line after an acute paracetamol (acetaminophen) overdose should be treated. This is a lower threshold than that used in the United States, Canada, Australia, and New Zealand. Here we report the impact of this change in the UK on the management of patients with acute overdose in different paracetamol concentration ranges.

METHODS

This is a cohort study, consisting of a retrospective analysis conducted on prospectively collected audit data in three UK hospitals. Following appropriate ethical and data protection authority approval, data for patients presenting within 24 h of an acute timed single paracetamol overdose were extracted. Numbers of admissions and use of antidote in relation to different paracetamol concentration bands (< 100 mg/L; 100-149 mg/L; 150-199 mg/L; and ≥ 200 mg/L at 4 h) were analyzed for one-year periods before and after the change.

RESULTS

Comparing the year before with the year after the change, there was no change in the numbers of patients presenting to hospital within 24 h of acute timed paracetamol overdose (1246 before and 1251 after), but more patients were admitted (759 before and 849 after) and treated with acetylcysteine (389 before and 539 after). Of the 150 additional patients treated with acetylcysteine in the year following the change, 114 (76%, 95% CI: 68.4-82.6) were in the 100-149 group and 9 (6.0%, 95% CI: 2.8-11.1) in the 150-199 group.

CONCLUSIONS

Changes to national guidelines for managing paracetamol poisoning in the UK have increased the numbers of patients with acute overdose treated with acetylcysteine, with most additional treatments occurring in patients in the 100-149 mg/L dose range, a group at low risk of hepatotoxicity and higher risk of adverse reactions.

The predictive value of hospital admission serum alanine transaminase activity in patients treated for paracetamol overdose

BACKGROUND

Paracetamol is a major cause of poisoning. Treatment decisions are predominately based on the dose ingested and a timed blood paracetamol concentration because most patients present to hospital soon after overdose, before hepatotoxicity can be confirmed/excluded using serum alanine transaminase (ALT). Nonetheless, ALT is measured at hospital presentation; we investigated its value in predicting hepatotoxicity.

METHODS

From March 2011 to May 2012, patients admitted to the Royal Infirmary of Edinburgh for paracetamol overdose treatment were identified. We determined the value of admission ALT (below or above our upper limit of normal-50 IU/l) at predicting three endpoints: 1-doubling of ALT; 2-peak ALT >1000 IU/l; 3-peak international normalized ratio (INR) >2.

RESULTS

From 500 patients, 410 met the entry criteria; 264 presented within 8 h of overdose, 54 between 8 and 24 h, 53 after 24 h and 39 were staggered ingestions. Admission ALT was increased in 71. For endpoint 1 (ALT doubling), the positive predictive value (PPV) of admission ALT was 19% [95% confidence interval (CI) 12-30] with a negative predictive value (NPV) of 98% (95% CI 96-99); endpoint 2 (ALT >1000 IU/l: PPV 23% (95% CI 14-34) and NPV 100% (95% CI 99-100) and for endpoint 3 (INR >2): PPV 14% (95% CI 7-25) and NPV of 100% (95% CI 99-100). The NPV remained high when only late presenters were included.

CONCLUSION

Admission ALT within the normal range has a high NPV and could be used, alone or in combination with newer biomarkers, to identify lower risk patients at hospital presentation.

Lipopolysaccharides and plant responses to phytopathogenic bacteria

Abstract Treatment of the leaves of pepper (Capsicum annuum) cv. ECW10R with lipopolysaccharides (LPS) from both plant pathogenic and enteric bacteria alters several aspects of the plant response to subsequent inoculation with phytopathogenic xanthomonads. LPS pre-treatment prevents the hypersensitive reaction caused by strains of Xanthomonas campestris pv. vesicatoria carrying the avirulence gene avrBs1 (a gene-for-gene interaction) and by X. campestris pv. campestris (a non-host interaction). Associated with this effect are the earlier synthesis of feruloyl- and coumaroyl-tyramine, phenolic conjugates that are potentially antimicrobial, and alterations in the expression patterns of genes for some pathogenesis-related (PR) proteins. Similar effects on the timing of phenolic conjugate synthesis are also seen in the compatible interaction with X. campestris pv. vesicatoria, although the level of the response is lower. Recognition of LPS by plants may allow expression of resistance in the absence of catastrophic tissue damage. However phytopathogenic bacteria may have evolved mechanisms to suppress the effects of LPS (and of other non-specific bacterial elicitors) on plant cells.

The epidemiology of mumps in southern Alberta 1980-1982

The full spectrum of mumps in Southern Alberta was studied over the years 1980-1982. In the community, a random sample of 3,497 children was tested for prevalence of immunity to mumps. The 1,816 who were not immune were retested after one year. The incidence of new immunity was 16.5%. Most seroconverters had no illness suggestive of mumps. History of previous mumps or immunization were found to be unreliable predictors of serologic immune status. In the family study, information was obtained on 342 cases diagnosed as mumps by family physician recorders. The overall complication rate was 10.8% and the secondary attack rate within families was 11.7%. Hospital records of all 25 patients admitted in Calgary with mumps during the same period were analyzed. Estimated costs of unopposed mumps over 20 years compared with the estimated costs of vaccination showed that a vaccination program could give a benefit-cost ratio between 6.4 and 247. These studies show that mumps is a mild disease with relatively few serious complications or sequelae. Even so, there would be definite medical and economic benefit with immunization. A low-cost addition of mumps vaccine to the immunization program would be justified.