IDMIL: an alignment-free Interpretable Deep Multiple Instance Learning (MIL) for predicting disease from whole-metagenomic data.

MOTIVATION: The human body hosts more microbial organisms than human cells. Analysis of this microbial diversity provides key insight into the role played by these microorganisms on human health. Metagenomics is the collective DNA sequencing of coexisting microbial organisms in an environmental sample or a host. This has several applications in precision medicine, agriculture, environmental science and forensics. State-of-the-art predictive models for phenotype predictions from metagenomic data rely on alignments, assembly, extensive pruning, taxonomic profiling and reference sequence databases. These processes are time consuming and they do not consider novel microbial sequences when aligned with the reference genome, limiting the potential of whole metagenomics. We formulate the problem of predicting human disease from whole-metagenomic data using Multiple Instance Learning (MIL), a popular supervised learning paradigm. Our proposed alignment-free approach provides higher accuracy in prediction by harnessing the capability of deep convolutional neural network (CNN) within a MIL framework and provides interpretability via neural attention mechanism. RESULTS: The MIL formulation combined with the hierarchical feature extraction capability of deep-CNN provides significantly better predictive performance compared to popular existing approaches. The attention mechanism allows for the identification of groups of sequences that are likely to be correlated to diseases providing the much-needed interpretation. Our proposed approach does not rely on alignment, assembly and reference sequence databases; making it fast and scalable for large-scale metagenomic data. We evaluate our method on well-known large-scale metagenomic studies and show that our proposed approach outperforms comparative state-of-the-art methods for disease prediction. AVAILABILITY AND IMPLEMENTATION: https://github.com/mrahma23/IDMIL.

Occupational exposure of aldehydes resulting from the storage of wood pellets.

An exposure assessment was conducted to investigate the potential for harmful concentrations of airborne short chain aldehydes emitted from recently stored wood pellets. Wood pellets can emit a number of airborne aldehydes include acetaldehyde, formaldehyde, propionaldehyde, butyraldehyde, valeraldehyde, and hexanal. Exposure limits have been set for these compounds since they can result in significant irritation of the upper respiratory system at elevated concentrations. Formaldehyde is a recognized human carcinogen and acetaldehyde is an animal carcinogen. Thus, air sampling was performed in a wood pellet warehouse at a pellet mill, two residential homes with bulk wood pellet storage bins, and in controlled laboratory experiments to evaluate the risk to occupants. Using NIOSH method 2539, sampling was conducted in five locations in the warehouse from April-June 2016 when it contained varying quantities of bagged pellets as well as two homes with ten ton bulk storage bins. The aldehyde concentrations were found to increase with the amount of stored pellets. Airborne concentrations of formaldehyde were as high as 0.45 ppm in the warehouse exceeding the NIOSH REL-C, and ACGIH TLV-C occupational exposure limits (OELs). The concentrations of aldehydes measured in the residential bins were also elevated indicating emissions may raise indoor air quality concerns for occupants. While individual exposures are of concern the combined irritant effect of all the aldehydes is a further raise the concerns for building occupants. To minimize exposure and the risk of adverse health effects to a building's occupants in storage areas with large quantities of pellets, adequate ventilation must be designed into storage areas.

Assessing pain severity and treatment outcomes in patients with low back pain: A Structural equation modeling approach at the center for the rehabilitation of the Paralysed, Bangladesh.

OBJECTIVES: This study aims to observe the associated risk factors of lower back pain and the factors that increase the pain severity. So, the main objective of this research is to identify the factors which may cause the lower back pain and the causal effect on the pain severity and respective treatment. This study also tries to determine the demographical characteristics of the low back pain patients and determine the inter relationship of psychological health, work stress and treatment effect with the pain disability index. STUDY DESIGN: In this cross-sectional study, 200 patients with lower back pain were interviewed who were taking treatments from the physiotherapy department at the Center for the Rehabilitation of the Paralysed, Savar, Dhaka, Bangladesh. METHODS: A quantitative research model has been performed to observe the relationship between different causes of low back pain effects on the patients. Different statistical analysis including structural equation modeling have been performed to observe their pain severity and treatment effect. RESULTS: The study found 64% (128) of the total participants as male and 36% (72) as female among 200 patients of low back pain. The study also observed the highest portion of the patients belong to the age group 39 to 45 years (21.5%). On the basis of BMI, obese weight respondents were 26.5% (53), overweight respondents were 37% (74), normal weight respondents were 33% (66), and underweight respondents were only 3.5% (7). Here, sex, body mass index (BMI), living place and educational status have significant association with pain disability index (PDI). On the other hand, smoking tendency of patients has insignificant relationship (p>0.05) with pain disability index (PDI). The path coefficients of the structural equation model identified that all the null hypotheses of no significant relationship have been rejected for 5% level of significance. The hypothesis of psychological health is positively related to pain severity of a patient has an acceptable strength (ß = 0.745, p<0.001) and a positive direction. Another hypothesis (Psychological health is positively related to the treatment of a patient) shows an acceptable strength (ß = 0.401, p <0.001) and a positive direction. Work stress is also found to be positively related to pain severity of a patient with an acceptable strength (ß = 0.544, p < 0.001) and a positive direction. The hypothesis (Work stress is positively related to the treatment of a patient) has an acceptable strength (ß = 0.322, p< 0.05) and a positive direction. The hypothesis (pain severity is positively related to the treatment of patients) shows an acceptable strength (ß = 0.801, p < 0.001) and a positive direction. CONCLUSION: The research found out the psychological health situation and work stress of patients are significantly related with pain severity with acceptable strength. Also, Pain severity is significantly associated with treatment scheme intensity.

A novel method for patterning of poly(3,4-ethylenedioxythiophene) films using UV exposure-activated self-assembled monolayers.

We have investigated a novel method for patterning of (3, 4-ethylenedioxythiophene) PEDOT, which has involved a selective polymerization of PEDOT on an UV-activated Self-Assembled-Monolayer surface. OTS coated surface has been activated by UV exposure, and the UV-exposed area served as adsorption sites for FeCl3 oxidants, providing a selective deposition of PEDOT films on FeCl3 adsorbed area, and thus leading to the selective patterning of PEDOT films. UV irradiation time and mask pattern dimension are main contributors to patternability: UV irradiation through Cr-mask (3 microm design) lead to approximately 3-5 microm patterns of PEDOT films, depending on the UV exposure time. In addition, a scotch tape peel test revealed excellent adhesion property of PEDOT to SiO2. Consequently, this simple method can be applied to define deep submicron dimensions due to its ability of providing a direct transfer of mask patterns to the substrate.

High performance flexible organic thin film transistors (OTFTs) with octadecyltrichlorsilane/ Al2O3/poly(4-vinylphenol) multilayer insulators.

The incorporation of a thin, atomic layer deposited Al2O3 layer in between a spin-coated poly-4-vinyl phenol (PVP) organic layer and octadecyltrichlorsilane (OTS) in the multilayer gate dielectric for pentacene organic thin film transistors on a n(+)-Si substrate reduced the gate leakage current and thereby significantly enhanced the current on/off ratio up to 2.8 x 10(6). Addition of the OTS monolayer on the UV-treated Al2O3 improved the crystallinity of the pentacene layer, where the OTS/UV-treated Al2O3 surfaces increased their contact angles to 100 degrees. X-ray diffraction (XRD) analysis revealed a more intense (001) crystal reflectance of pentacene deposited on OTS/UV-treated Al2O3 surface than that on OTS/Al2O3 surface. Moreover, the improved pentacene layer contributed to the field effect mobility (0.4 cm2/Vs) and subsequently improved the electrical performances of organic thin film transistor (OTFT) devices. This PVP/UV treated Al2O3/OTS multilayer gate dielectric stack was superior to those of the device with the single PVP gate dielectrics due to the improved crystallinity of pentacene.

Phenotype Prediction from Metagenomic Data Using Clustering and Assembly with Multiple Instance Learning (CAMIL).

The recent advent of Metagenome Wide Association Studies (MGWAS) provides insight into the role of microbes on human health and disease. However, the studies present several computational challenges. In this paper, we demonstrate a novel, efficient, and effective Multiple Instance Learning (MIL) based computational pipeline to predict patient phenotype from metagenomic data. MIL methods have the advantage that besides predicting the clinical phenotype, we can infer the instance level label or role of microbial sequence reads in the specific disease. Specifically, we use a Bag of Words method, which has been shown to be one of the most effective and efficient MIL methods. This involves assembly of the metagenomic sequence data, clustering of the assembled contigs, extracting features from the contigs, and using an SVM classifier to predict patient labels and identify the most relevant sequence clusters. With the exception of the given labels for the patients, this entire process is de novo (unsupervised). We call our pipeline "CAMIL", which stands for Clustering and Assembly with Multiple Instance Learning. We use multiple state-of-the-art clustering methods for feature extraction, evaluation, and comparison of the performance of our proposed approach for each of these clustering methods. We also present a fast and scalable pre-clustering algorithm as a preprocessing step for our proposed pipeline. Our approach achieves efficiency by partitioning the large number of sequence reads into groups (called canopies) using locality sensitive hashing (LSH). These canopies are then refined by using state-of-the-art sequence clustering algorithms. We use data from a well-known MGWAS study of patients with Type-2 Diabetes and show that our pipeline significantly outperforms the classifier used in that paper, as well as other common MIL methods.

Analysis of Historical Worker Exposures to Respirable Dust from Talc Mining and Milling Operations in Vermont.

OBJECTIVES: Talc is mined and milled throughout the world for use in a variety of industrial and consumer products. Although prior studies have evaluated workplace exposures or health effects from talc operations, the primary emphasis of these investigations has been on certain mineral contaminants (e.g. crystalline silica and asbestos) rather than talc itself. The purpose of this analysis is to evaluate historical worker exposures to respirable dust (as a measure of talc exposures) in the Vermont talc mines and mills, which involved a relatively pure form of talc (i.e. no asbestos and <0.25% or <1% crystalline silica). METHODS: Respirable dust sampling data collected for workers in the Vermont mines and mills, which have not been previously published, were obtained from both mining company records and Mine Safety and Health Administration (MSHA) inspections. Because of differences in sampling design, the company and MSHA data were analyzed and reported separately. Overall, nearly 700 respirable dust samples collected for 44 job categories at 7 site locations over an approximate 30-year period were analyzed. RESULTS: Average respirable dust concentrations were found to exceed occupational exposure limits (OELs) in the United States and other countries for several job categories and site locations. Regardless of data source, the highest observed exposures were for mining jobs involving the operation of heavy equipment to break up, move, or load raw ore from the mines and milling or shipping jobs involving the crushing of raw ore, cleaning and drying of processed talc, and bagging and packaging of the final talc product. When analyzing the company data, the arithmetic mean respirable dust concentration was 2.73 mg m-3 for Muckerman at Hammondsville Mine, 3.18 mg m-3 for dosco operator at Ludlow mines, 1.35 mg m-3 for crusher operator at Gassetts Mill, 2.4 mg m-3 for palletizer at West Windsor Mill, and 2.68 mg m-3 for bagging operator at Columbia Shipping Center. When analyzing the MSHA data, the arithmetic mean respirable dust concentration was 3.5 mg m-3 for kiln/dryer operator at Hammondsville Mine, 1.27 mg m-3 for driller at Ludlow mines, 3.69 mg m-3 for ball mill operator at Columbia mill, 3.02 mg m-3 for flotation operator at West Windsor Mill, and 3.24 mg m-3 for bagging operator at Columbia Shipping Center. Worker exposures were found to decline over time for many, but not all, jobs. CONCLUSIONS: Our findings highlight potential high-risk jobs that might benefit from additional exposure control strategies at current or future talc manufacturing sites. The respirable dust measurements summarized here may also be used to reconstruct historical worker exposures at the Vermont sites or aid in subsequent epidemiology studies of this cohort focused on malignant or non-malignant respiratory disease.

Current State of and Future Opportunities for Prediction in Microbiome Research: Report from the Mid-Atlantic Microbiome Meet-up in Baltimore on 9 January 2019.

Accurate predictions across multiple fields of microbiome research have far-reaching benefits to society, but there are few widely accepted quantitative tools to make accurate predictions about microbial communities and their functions. More discussion is needed about the current state of microbiome analysis and the tools required to overcome the hurdles preventing development and implementation of predictive analyses. We summarize the ideas generated by participants of the Mid-Atlantic Microbiome Meet-up in January 2019. While it was clear from the presentations that most fields have advanced beyond simple associative and descriptive analyses, most fields lack essential elements needed for the development and application of accurate microbiome predictions. Participants stressed the need for standardization, reproducibility, and accessibility of quantitative tools as key to advancing predictions in microbiome analysis. We highlight hurdles that participants identified and propose directions for future efforts that will advance the use of prediction in microbiome research.

Carbon Monoxide Off-Gassing From Bags of Wood Pellets.

Wood pellets are increasingly used for space heating in the United States and globally. Prior work has shown that stored bulk wood pellets produce sufficient carbon monoxide (CO) to represent a health concern and exceed regulatory standards for occupational exposures. However, most of the pellets used for residential heating are sold in 40-pound (18.1 kg) plastic bags. This study measured CO emission factors from fresh, bagged-wood pellets as a function of temperature and relative humidity. CO concentrations increased with increasing temperature and moisture in the container. CO measurements in a pellet mill warehouse with stored pallets of bagged pellets had 8-h average CO concentrations up to 100 ppm exceeding occupational standards for worker exposure. Thus, manufacturers, distributors, and home owners should be aware of the potential for CO in storage areas and design facilities with appropriate ventilation and CO sensors.

Chemical characterization of PM2.5 collected from a rural coastal island of the Bay of Bengal (Bhola, Bangladesh).

This work focuses on the chemical characterization of fine aerosol particles (PM2.5) collected from a rural remote island of the Bay of Bengal (Bhola, Bangladesh) from April to August, 2013. PM2.5 particle-loaded filters were analyzed for organic carbon (OC), elemental carbon (EC), water-soluble ions, and selected saccharides (levoglucosan, mannosan, galactosan, arabitol, and mannitol). The average PM2.5 mass was 15.0 ± 6.9 µg m-3. Organic carbon and elemental carbon comprised roughly half of the analyzed components. Organic carbon was the predominant contributor to total carbon (TC) and accounting for about 28% of PM2.5 mass. Secondary organic carbon (SOC) was inferred to be ~ 26% of OC. The sum of ions comprised ~ 27% of PM2.5 mass. The contribution of sea salt aerosol was smaller than expected for a sea-near site (17%), and very high chloride depletion was observed (78%). NssSO42- was a dominant ionic component with an average concentration of 2.0 µg m-3 followed by Na+, NH4+, and nssCa2+. The average concentration of arabitol and mannitol was 0.11 and 0.14 µg m-3, respectively, while levoglucosan and its stereoisomers (mannosan and galactosan) were bellow detection limit. NH4+/SO42- equivalent ratio was 0.30 ± 0.13 indicating that secondary inorganic aerosol is not the main source of SO42-. Enrichment factor (EF) analysis showed that SO42- and NO3- were enriched in atmospheric particles compared to sea aerosol and soil indicating their anthropogenic origin. Higher OC/EC ratio (3.70 ± 0.88) was a good indicator of the secondary organic compounds formation. Other ratios (OC/EC, K+/EC, nssSO42-/EC) and correlation analysis suggested mixed sources for carbonaceous components. Arabitol and mannitol both showed strong correlation with EC having R 2 value 0.89 and 0.95, respectively. Air mass trajectories analysis showed that concentrations of soil and anthropogenic species were lower for air masses originating from the sea (May-August) and were higher when air came from land (April).

Metagenome sequence clustering with hash-based canopies.

Metagenomics is the collective sequencing of co-existing microbial communities which are ubiquitous across various clinical and ecological environments. Due to the large volume and random short sequences (reads) obtained from community sequences, analysis of diversity, abundance and functions of different organisms within these communities are challenging tasks. We present a fast and scalable clustering algorithm for analyzing large-scale metagenome sequence data. Our approach achieves efficiency by partitioning the large number of sequence reads into groups (called canopies) using hashing. These canopies are then refined by using state-of-the-art sequence clustering algorithms. This canopy-clustering (CC) algorithm can be used as a pre-processing phase for computationally expensive clustering algorithms. We use and compare three hashing schemes for canopy construction with five popular and state-of-the-art sequence clustering methods. We evaluate our clustering algorithm on synthetic and real-world 16S and whole metagenome benchmarks. We demonstrate the ability of our proposed approach to determine meaningful Operational Taxonomic Units (OTU) and observe significant speedup with regards to run time when compared to different clustering algorithms. We also make our source code publicly available on Github. a.

Slurry sampling techniques for the determination of lead in Bangladeshi fish samples by electrothermal atomic absorption spectrometry with a metal tube atomizer.

Ultrasonic slurry sampling electrothermal atomic absorption spectrometry with a metal tube atomizer has been applied to the determination of lead in Bangladeshi fish samples. The slurry sampling conditions, such as slurry stabilizing agent, slurry concentration, pyrolysis temperature for the slurried fish samples, particle size and ultrasonic agitation time, were optimized for electrothermal atomic absorption spectrometry with the Mo tube atomizer. Thiourea was used as the chemical modifier for the interference of matrix elements. The detection limit was 53 fg (3S/N). The determined amount of lead in Bangladeshi fish samples was consistent with those measured in the dissolved acid-digested samples. The advantages of the proposed methods are easy calibration, simplicity, low cost and rapid analysis.

Reversed Phase SPE and GC-MS Study of Polycyclic Aromatic Hydrocarbons in Water Samples from the River Buriganga, Bangladesh.

Polycyclic aromatic hydrocarbons (PAHs) are semivolatile organic compounds (SVOCs) categorized as persistent organic pollutants (POPs). PAHs are ubiquitous in terrestrial, atmospheric, and particularly aquatic environments throughout the world and have been detected in lakes, ground waters, and rivers. This research work involved the analysis of five PAHs, anthracene, fluorene, naphthalene, phenanthrene, and pyrene, in water sample collected from the river Buriganga, Bangladesh. The extraction of water samples was carried out by reversed phase solid-phase extraction (RP-SPE) technique with C-18 SPE cartridges. A solvent mixture of dichloromethane and hexane (1 : 2) with a flow rate of 0.5 mL/min was used as eluent. Percentage recoveries of five PAHs for this technique were in the range of 81.47 ± 1.16 to 98.60 ± 0.61%. PAHs quantification was achieved by using an ion trap gas chromatography mass spectrometer (GC-MS) interfaced to gas chromatography (GC) equipped with a fused silica capillary column. Helium was used as carrier gas with a flow rate of 1.0 mL/min. The commonly detected PAH compounds in the river water were anthracene, naphthalene, and phenanthrene at the concentration ranges of 0.451 to 3.201, 0.033 to 3.1131, and 0.320 to 2.546 µg/mL, respectively. The results reflect that PAHs presented in this river water were mostly from petrogenic and pyrogenic sources.

Comparative analysis of serum manganese, zinc, calcium, copper and magnesium level in panic disorder patients.

The purpose of the study was to determine the serum concentration of trace elements of panic disorder patients and to find out the relationship between trace element levels and nutritional status or socio-economic factors. The study was conducted among 54 panic disorder patients and 52 healthy volunteers. Patients were recruited from Bangabandhu Sheikh Mujib Medical University by random sampling. Serum trace element concentrations were determined by flame atomic absorption spectroscopy (for Mg, Zn, Ca, and Cu) as well as graphite furnace (for Mn). Data were analyzed by independent t test, Pearson's correlation analysis, regression analysis, and ANOVA. The serum concentration of Mn, Zn, Ca, Cu, and Mg in panic disorder patients were 0.37 +/- 0.30, 0.67 +/- 0.20, 99.91 +/- 15.15, 0.83 +/- 0.23, and 21.14 +/- 3.72 mg/L, while those were 0.4163 +/- 0.2527, 0.86 +/- 0.3, 106.6073 +/- 18.6531, 0.8514 +/- 0.3646, and 21.37 +/- 2.03 mg/L in control subjects, respectively. The serum concentration of Zn decreased significantly (p = 0.001) in patient group. But the differences of the concentration of Mn, Ca, Cu, and Mg between patient and control group were not significant (p = 0.522, p = 0.065, p = 0.800, and p = 0.712, respectively). Socio-economic data reveal that most of the patients were very poor and middle aged. Mean BMIs of the control group (23.74 +/- 2.71 kg/m(2)) and the patient group (22.62 +/- 3.74 kg/m(2)) were within the normal range (18.5-25.0 kg/m(2)). There was no significant relationship between serum zinc level and BMI of patients (r = 0.038; p = 0.809). So the decreased level of serum zinc in panic disorder patients was not because of other reasons, but rather it may provide a prognostic tool for the diagnosis and treatment of this disease.

Wastewater treatment with multilayer media of waste and natural indigenous materials.

Wastewater treatment using waste materials (refuse concrete, waste paper and charcoal) and natural indigenous rocks (andesite, limestone, granite and nitrolite) in the form of multilayer media was investigated. The removal of suspended solids (SS), phosphate ion, nitrate ion, ammonium ion, toxic metals and chemical oxygen demand (COD) were evaluated for the multilayer wastewater treatment system. Effective removal of heavy metals such as cadmium, chromium, mercury and lead was demonstrated. SS and phosphate ion were removed with relatively high efficiency and the COD after treatment was lessened using certain combinations of media. The present wastewater treatment system is simple, convenient and low cost. Therefore, this method can be applied in small scale plants for wastewater treatment in local and nonexclusive areas.

Determination of silver in environmental samples by tungsten wire preconcentration method-electrothermal atomic absorption spectrometry.

A preconcentration method for silver in environmental waters involving adsorption on a tungsten wire, followed by electrothermal atomic absorption spectrometry with a tungsten tube atomizer is described. The optimal immersing time was 90s. The best pH for the adsorption of silver was 3. Under the optimal conditions, the detection limit for silver by the tungsten wire preconcentration method was 5.0ngl(-1) (3S/N) and the relative standard deviation was 8.2%. The effects of large amounts of concomitants on the preconcentration of silver were evaluated. Even though 10(3)- to 10(4)-fold excess of matrix elements existed in water, the silver response was not significantly affected by the matrix elements. The method with preconcentration on a tungsten wire was applied to the determination of silver in waters and proved to be sensitive, simple, and convenient. This adsorption method can be utilized in in situ sampling of ultra-trace silver in environmental samples (waters). Furthermore, after sampling it is easy to carry and store the tungsten wire without contamination for a long time.

Separation of zinc compounds by sequential metal vapor elution analysis with atomic absorption detection.

The separation of zinc compounds, containing zinc chloride, nitrate, and sulfate, at low concentrations by sequential metal vapor elution analysis (SMVEA) with argon carrier gas was reported. A molybdenum column, inserted with a tungsten wire, was developed for the separation of zinc compounds by SMVEA. The optimum separation conditions were a vaporization temperature of 1370K, a column temperature of 1350K, and a carrier gas flow rate of 2.5mLmin(-1). Under the optimized experimental conditions, the zinc compounds could be roughly separated by SMVEA, although a part of peak profiles overlapped. The number of theoretical plates was 36 for ZnCl(2), 62 for Zn(NO(3))(2), and 80 for ZnSO(4) in the SMVEA column. The present SMVEA system may be able to be applied widely to various analytical instruments.