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

General General

Predictive modeling of pharmaceutical product removal by a managed aquifer recharge system: Comparison and optimization of models using ensemble learners.

In Journal of environmental management

Pharmaceutical products (PPs) are emerging water pollutants with adverse environmental and health-related impacts, owing to their toxic, persistent, and undetectable microscopic nature. Globally, increasing scientific knowledge and advanced technologies have allowed researchers to study PP-associated problems and their removal for water reuse. Experimental modeling methods require laborious, lengthy, expensive, and environmentally hazardous lab-work to optimize the process. On the other hand, predictive machine learning (ML) models can trace the complex input-output relationship of a process using available datasets. In this study, ensemble ML techniques, including decision tree (DT), random forest (RF), and Xtreme gradient boost (XGB), were used to explore PP (diclofenac, iopromide, propranolol, and trimethoprim) removal by a managed aquifer recharge (MAR) system. The model input parameters included characteristics of reclaimed water and soil used in the columns, pH, dissolved organic carbon, operating time, nitrogen dioxide, sulfate, nitrate, electrical conductivity, manganese, and iron. The selected PP removal was the model output. Datasets were collected through a one-year experimental study of continuous MAR system operation to predict the removal of PPs. DT, RF, and XGB models were then developed for one of the selected compounds and tested for the others to check the reliability of the ML model results. The developed models were assessed using statistical performance matrices. The experimental results showed >80% removal of propranolol and trimethoprim; however, removal of diclofenac and iopromide was only ≈50% by the MAR system. The proposed DT and RF models presented higher coefficients of determination (R2 ≥ 0.92) for diclofenac, propranolol, and trimethoprim than for iopromide (R2 ≤ 0.63). In contrast, the XGB model showed better results for diclofenac, iopromide, propranolol, and trimethoprim, with R2 values of 0.92, 0.72, 0.96, and 0.97, respectively. Therefore, XGB could be the best predictive model to provide insight into the adaptation of ML models to predict PP removal by the MAR system, thereby minimizing experimental work.

Yaqub Muhammad, Ngoc Nguyen Mai, Park Soohyung, Lee Wontae

2022-Sep-30

Decision tree, Managed aquifer recharge system, Pharmaceutical, Random forest, Xtreme gradient boost

Public Health Public Health

Biomarkers of maternal lead exposure during pregnancy using micro-spatial child deciduous dentine measurements.

In Environment international

BACKGROUND : Lead is a toxic chemical of public health concern, however limited biomarkers are able to reconstruct prior lead exposures in early-life when biospecimens are not collected and stored. Although child tooth dentine measurements accurately assess past child perinatal lead exposure, it has not been established if they reflect maternal exposure in pregnancy.

AIM : To assess the prenatal relationship between child tooth dentine and maternal blood lead measurements and to estimate maternal lead exposure during the 2nd and 3rd trimesters of pregnancy from weekly child dentine profiles.

METHODS : We measured early-life lead exposure in child tooth dentine and maternal blood from 419 child-mother dyads enrolled in the Programming Research in Obesity, Growth, Environment and Social Stress (PROGRESS) cohort. We employed the Super-Learner algorithm to determine the relationship of dentine lead data with maternal blood lead concentrations and to predict maternal lead from child dentine lead data in blinded analyses. We validated and quantified the bias of our results internally.

RESULTS : Mothers had moderate blood lead levels (trimesters: 2nd = 29.45 ug/L, 3rd = 31.78 ug/L). Trimester-averaged and weekly child dentine lead measurements were highly correlated with maternal blood levels in the corresponding trimesters. The predicted trimester-specific maternal lead levels were significantly correlated with actual measured blood values (trimesters: 2nd = 0.83; 3rd = 0.88). Biomarkers of maternal lead exposure discriminated women highly exposed to lead (>mean) with 85 % and 96 % specificity in the 2nd and 3rd trimesters, respectively, with 80 % sensitivity.

DISCUSSION : Weekly child dentine lead levels can serve as biomarkers of past child and maternal lead exposures during pregnancy.

Gerbi Lucia, Austin Christine, Pedretti Nicolo Foppa, McRae Nia, Amarasiriwardena Chitra J, Mercado-García Adriana, Torres-Olascoaga Libni A, Tellez-Rojo Martha M, Wright Robert O, Arora Manish, Elena Colicino

2022-Sep-16

Blood lead levels, Machine learning, Pregnancy, Prenatal lead exposure, Super-Learner algorithm, Tooth dentine lead levels

General General

Development and validation of a microenvironment-related prognostic model for hepatocellular carcinoma patients based on histone deacetylase family.

In Translational oncology

BACKGROUND : Histone deacetylase (HDAC) family can remove acetyl groups from histone lysine residues, and their high expression is closely related to the poor prognosis of hepatocellular carcinoma (HCC) patients. Recently, it has been reported to play an immunosuppressive role in the microenvironment, but little is known about the mechanism.

METHODS : Through machine learning, we trained and verified the prognostic model composed of HDACs. CIBERSORT was used to calculate the percentage of immune cells in the microenvironment. Based on co-expression network, potential targets of HDACs were screened. After that, qRT-PCR was employed to evaluate the expression of downstream genes of HDACs, while HPLC-CAD analysis was applied to detect the concentration of arachidonic acid (AA). Finally, Flow cytometry, WB and IHC experiments were used to detect CD86 expression in RAW246.7.

RESULTS : We constructed a great prognostic model composed of HDAC1 and HDAC11 that was significantly associated with overall survival. These HDACs were related to the abundance of macrophages, which might be attributed to their regulation of fatty-acid-metabolism related genes. In vitro experiments, the mRNA expression of ACSM2A, ADH1B, CYP2C8, CYP4F2 and SLC27A5 in HCC-LM3 was significantly down-regulated, and specific inhibitors of HDAC1 and HDAC11 significantly promoted the expression of these genes. HDAC inhibitors can promote the metabolism of AA, which may relieve the effect of AA on the polarization of M1 macrophages.

CONCLUSIONS : Our study revealed the blocking effect of HDAC1 and HDAC11 on the polarization of macrophages M1 in the microenvironment by inhibiting fatty acid metabolism.

Teng Linxin, Li Zhengjun, Shi Yipeng, Gao Zihan, Yang Yang, Wang Yunshan, Bi Lei

2022-Sep-30

Fatty acid metabolism, Hepatocellular carcinoma, Histone deacetylase, Macrophages, Tumor microenvironment

General General

Real-time driving risk assessment using deep learning with XGBoost.

In Accident; analysis and prevention

Traffic crashes typically occur in a few seconds and real-time prediction can significantly benefit traffic safety management and the development of safety countermeasures. This paper presents a novel deep learning model for crash identification based on high-frequency, high-resolution continuous driving data. The method consists of feature engineering based on Convolutional Neural Network (CNN) and Gated Recurrent Unit (GRU) and classification based on Extreme Gradient Boosting (XGBoost). The CNN-GRU architecture captures the time series characteristics of driving kinematics data. Compared to normal driving segments, safety-critical events (SCEs)-i.e., crashes and near-crashes (CNC)-are rare. The weighted categorical cross-entropy loss and oversampling methods are utilized to address this imbalance issue. An XGBoost classifier is utilized instead of the multi-layer perceptron (MLP) to achieve a high precision and recall rate. The proposed approach is applied to the Second Strategic Highway Research Program Naturalistic Driving Study (SHRP 2 NDS) data with 1,820 crashes, 6,848 near-crashes, and 59,997 normal driving segments. The results show that in a 3-class classification system (crash, near-crash, normal driving segments), the accuracy for the overall model is 97.5%, and the precision and recall for crashes are 84.7%, and 71.3% respectively, which is substantially better than benchmarks models. Furthermore, the recall of the most severe crashes is 98.0%. The proposed crash identification approach provides an accurate, highly efficient, and scalable way to identify crashes based on high frequency, high-resolution continuous driving data and has broad application prospects in traffic safety applications.

Shi Liang, Qian Chen, Guo Feng

2022-Sep-30

Convolutional neural network, Crash prediction, Deep learning, Gated recurrent unit, High frequency kinematic driving data, Naturalistic driving study, XGBoost

General General

Quantization-aware training for low precision photonic neural networks.

In Neural networks : the official journal of the International Neural Network Society

Recent advances in Deep Learning (DL) fueled the interest in developing neuromorphic hardware accelerators that can improve the computational speed and energy efficiency of existing accelerators. Among the most promising research directions towards this is photonic neuromorphic architectures, which can achieve femtojoule per MAC efficiencies. Despite the benefits that arise from the use of neuromorphic architectures, a significant bottleneck is the use of expensive high-speed and precision analog-to-digital (ADCs) and digital-to-analog conversion modules (DACs) required to transfer the electrical signals, originating from the various Artificial Neural Networks (ANNs) operations (inputs, weights, etc.) in the photonic optical engines. The main contribution of this paper is to study quantization phenomena in photonic models, induced by DACs/ADCs, as an additional noise/uncertainty source and to provide a photonics-compliant framework for training photonic DL models with limited precision, allowing for reducing the need for expensive high precision DACs/ADCs. The effectiveness of the proposed method is demonstrated using different architectures, ranging from fully connected and convolutional networks to recurrent architectures, following recent advances in photonic DL.

Kirtas M, Oikonomou A, Passalis N, Mourgias-Alexandris G, Moralis-Pegios M, Pleros N, Tefas A

2022-Sep-19

Constrained-aware training, Neural network quantization, Photonic deep learning

General General

Raman spectroscopy and FTIR spectroscopy fusion technology combined with deep learning: A novel cancer prediction method.

In Spectrochimica acta. Part A, Molecular and biomolecular spectroscopy

According to the limited molecular information reflected by single spectroscopy, and the complementarity of FTIR spectroscopy and Raman spectroscopy, we propose a novel diagnostic technology combining multispectral fusion and deep learning. We used serum samples from 45 healthy controls, 44 non-small cell lung cancer (NSCLC), 38 glioma and 37 esophageal cancer patients, and the Raman spectra and FTIR spectra were collected respectively. Then we performed low-level fusion and feature fusion on the spectral, and used SVM, Convolutional Neural Network-Long-Short Term Memory (CNN-LSTM) and the multi-scale convolutional fusion neural network (MFCNN). The accuracy of low-level fusion and feature fusion models are improved by about 10% compared with single spectral models.

Leng Hongyong, Chen Cheng, Chen Chen, Chen Fangfang, Du Zijun, Chen Jiajia, Yang Bo, Zuo Enguang, Xiao Meng, Lv Xiaoyi, Liu Pei

2022-Sep-20

FTIR spectroscopy, Feature fusion, Low-level fusion, MFCNN, Raman