Optimizing the classification of biological tissues using machine learning models based on polarized data

Polarimetric data is nowadays used to build recognition models for the characterization of organic tissues or the early detection of some diseases. Different Mueller matrix-derived polarimetric observables, which allow a physical interpretation of a specific characteristic of samples, are proposed in literature to feed the required recognition algorithms. However, they are obtained through mathematical transformations of the Mueller matrix and this process may loss relevant sample information in search of physical interpretation. In this work, we present a thorough comparative between 12 classification models based on different polarimetric datasets to find the ideal polarimetric framework to construct tissues classification models. The study is conducted on the experimental Mueller matrices images measured on different tissues: muscle, tendon, myotendinous junction and bone; from a collection of 165 ex-vivo chicken thighs. Three polarimetric datasets are analyzed: (A) a selection of most representative metrics presented in literature; (B) Mueller matrix elements; and (C) the combination of (A) and (B) sets. Results highlight the importance of using raw Mueller matrix elements for the design of classification models.     

Characterization of ionic strength for X-MuLV inactivation by low pH treatment for monoclonal antibody purification

In the production of monoclonal antibodies (mAbs) intended for use in humans, it is a global regulatory requirement that the manufacturing process includes unit operations that are proven to inactivate or remove adventitious agents to ensure viral safety. Viral inactivation by low pH hold (LPH) is typically used to ensure this viral safety in the purification process of mAbs and other biotherapeutics derived from mammalian cell lines. To ascertain the effectiveness of the LPH step, viral clearance studies have evaluated LPH under worst-case conditions of pH above the manufacturing set point and hold duration at or below the manufacturing minimum. Highly acidic conditions (i.e., pH < 3.60) provide robust and effective enveloped virus inactivation but may lead to reduced product quality of the therapeutic protein. However, when viral inactivation is operated above pH 3.60 to ensure product stability, effective (>4 log₁₀ reduction factor) viral inactivation may not be observed under these worst-case pH conditions in viral clearance studies. A multivariate design of experiments was conducted to further characterize the operating space for low pH viral inactivation of a model retrovirus, xenotropic murine leukemia virus (X-MuLV). The statistically designed experiment evaluated the effect of mAb isotype, pH, temperature, acid titrant, sodium chloride (NaCl) concentration, virus spike timing, and post-spike filtration on X-MuLV inactivation. Data from the characterization study were used to generate predictive models to identify conditions that reliably achieve effective viral inactivation at pH ≥ 3.60. Results of the study demonstrated that NaCl concentration has the greatest effect on virus inactivation in the range studied, and pH has a large effect when the load material has no additional NaCl. Overall, robust and effective inactivation of X-MuLV at pH 3.65–3.80 can be achieved by manipulating either the pH or the NaCl concentration of the load material. This study contributes to the understanding of ionic strength as an influential parameter in low pH viral inactivation studies.     

Suitability of a generic virus safety evaluation for monoclonal antibody investigational new drug applications

Biologics produced from CHO cell lines with endogenous virus DNA can produce retrovirus-like particles in cell culture at high titers, and other adventitious viruses can find their way through raw materials into the process to make a product. Therefore, it is the industry standard to have controls to avoid introduction of viruses into the production process, to test for the presence of viral particles in unclarified cell culture, and to develop purification procedures to ensure that manufacturing processes are robust for viral clearance. Data have been accumulated over the past four decades on unit operations that can inactivate and clear adventitious virus and provide a high degree of assurance for patient safety. During clinical development, biological products are traditionally tested at process set points for viral clearance. However, the widespread implementation of platform production processes to produce highly similar IgG antibodies for many indications makes it possible to leverage historical data and knowledge from representative molecules to allow for better understanding and control of virus safety. More recently, individualized viral clearance studies are becoming the rate-limiting step in getting new antibody molecules to clinic, particularly in Phase 0 and eIND situations. Here, we explore considerations for application of a generic platform virus clearance strategy that can be applied for relevant investigational antibodies within defined operational parameters in order to increase speed to the clinic and reduce validation costs while providing a better understanding and assurance of process virus safety.     

机器学习在MALDI-TOF MS鉴定微生物中的应用

基质辅助激光解吸/电离飞行时间质谱(matrix-assisted laser desorption/ionization time-of-flight mass spectrometry,MALDI-TOF MS)是一种新兴的高通量技术,已广泛应用于临床微生物、食品微生物和水产微生物的快速鉴定。如何进一步提高MALDI-TOF MS在微生物鉴定中的分辨率是该技术当前面临的一大挑战。为了高效处理大量高维微生物MALDI-TOF MS数据,各种机器学习算法得到了应用。本文综述了机器学习在微生物MALDI-TOFMS鉴定中的应用。首先,本文在介绍机器学习在微生物MALDI-TOF MS分类中的工作流程后,进一步对MALDI-TOF MS的数据特征、MALDI-TOF MS数据库、数据的预处理和模型的性能评估进行了描述。然后讨论了典型的机器学习分类算法和集成学习算法的应用。简单的机器学习算法很难满足微生物MALDI-TOF MS分类的高分辨率的需求,而组合不同机器学习算法和集成学习算法可以获得更好的微生物分类性能。在MALDI-TOF MS数据的预处理方面,小波算法和遗传算法的应用最广,它们...     

Prediction of effluent quality in papermaking wastewater treatment processes using dynamic kernel-based extreme learning machine

Papermaking wastewater accounts for a large proportion of industrial wastewater, and it is essential to obtain accurate and reliable effluent indices in real-time. Considering the complexity, nonlinearity, and time variability of wastewater treatment processes, a dynamic kernel extreme learning machine (DKELM) method is proposed to predict the key quality indices of effluent chemical oxygen demand (COD). A time lag coefficient is introduced and a kernel function is embedded into the extreme learning machine (ELM) to extract dynamic information and obtain better prediction accuracy. A case study for modeling a wastewater treatment process is demonstrated to evaluate the performance of the proposed DKELM. The results illustrate that both training and prediction accuracy of the DKELM model is superior to other models. For the prediction of the quality indices of effluent COD, the determinate coefficient of the DKELM model is increased by 27.52 %, 21.36 %, 10.42 %, and 10.81 %, compared with partial least squares, ELM, dynamic ELM, and kernel ELM, respectively.     

Anion exchange chromatography provides a robust, predictable process to ensure viral safety of biotechnology products

The mammalian cell-lines used to produce biopharmaceutical products are known to produce endogenous retrovirus-like particles and have the potential to foster adventitious viruses as well. To ensure product safety and regulatory compliance, recovery processes must be capable of removing or inactivating any viral impurities or contaminants which may be present. Anion exchange chromatography (AEX) is a common process in the recovery of monoclonal antibody products and has been shown to be effective for viral removal. To further characterize the robustness of viral clearance by AEX with respect to process variations, we have investigated the ability of an AEX process to remove three model viruses using various combinations of mAb products, feedstock conductivities and compositions, equilibration buffers, and pooling criteria. Our data indicate that AEX provides complete or near-complete removal of all three model viruses over a wide range of process conditions, including those typically used in manufacturing processes. Furthermore, this process provides effective viral clearance for different mAb products, using a variety of feedstocks, equilibration buffers, and different pooling criteria. Viral clearance is observed to decrease when feedstocks with sufficiently high conductivities are used, and the limit at which the decrease occurs is dependent on the salt composition of the feedstock. These data illustrate the robust nature of the AEX recovery process for removal of viruses, and they indicate that proper design of AEX processes can ensure viral safety of mAb products.     

Application of radiomics and machine learning in head and neck cancers

With the continuous development of medical image informatics technology, more and more high-throughput quantitative data could be extracted from digital medical images, which has resulted in a new kind of omics-Radiomics. In recent years, in addition to genomics, proteomics and metabolomics, radiomic has attracted the interest of more and more researchers. Compared to other omics, radiomics can be perfectly integrated with clinical data, even with the pathology and molecular biomarker, so that the study can be closer to the clinical reality and more revealing of the tumor development. Mass data will also be generated in this process. Machine learning, due to its own characteristics, has a unique advantage in processing massive radiomic data. By analyzing mass amounts of data with strong clinical relevance, people can construct models that more accurately reflect tumor development and progression, thereby providing the possibility of personalized and sequential treatment of patients. As one of the cancer types whose treatment and diagnosis rely on imaging examination, radiomics has a very broad application prospect in head and neck cancers (HNC). Until now, there have been some notable results in HNC. In this review, we will introduce the concepts and workflow of radiomics and machine learning and their current applications in head and neck cancers, as well as the directions and applications of artificial intelligence in the treatment and diagnosis of HNC.     

机器学习方法在CRISPR/Cas9系统中的应用

基于CRISPR/Cas9系统介导的第三代基因组定点编辑技术,已被广泛应用于基因编辑和基因表达调控等研究领域。如何提高该技术对基因组编辑的效率与特异性、最大限度降低脱靶风险一直是该领域的难点。近年来,机器学习为解决CRISPR/Cas9系统所面临的问题提供了新思路,基于机器学习的CRISPR/Cas9系统已逐渐成为研究热点。本文阐述了CRISPR/Cas9的作用机理,总结了现阶段该技术面临的基因组编辑效率低、存在潜在的脱靶效应、前间区序列邻近基序(PAM)限制识别序列等问题,最后对机器学习应用于优化设计高效向导RNA (sgRNA)序列、预测sgRNA的活性、脱靶效应评估、基因敲除、高通量功能基因筛选等领域的研究现状与发展前景进行了展望,以期为基因组编辑领域的研究提供参考。     

Global metabolic profiling to model biological processes of aging in twins

Aging is intimately linked to system‐wide metabolic changes that can be captured in blood. Understanding biological processes of aging in humans could help maintain a healthy aging trajectory and promote longevity. We performed untargeted plasma metabolomics quantifying 770 metabolites on a cross‐sectional cohort of 268 healthy individuals including 125 twin pairs covering human lifespan (from 6 months to 82 years). Unsupervised clustering of metabolic profiles revealed 6 main aging trajectories throughout life that were associated with key metabolic pathways such as progestin steroids, xanthine metabolism, and long‐chain fatty acids. A random forest (RF) model was successful to predict age in adult subjects (≥16 years) using 52 metabolites (R² = .97). Another RF model selected 54 metabolites to classify pediatric and adult participants (out‐of‐bag error = 8.58%). These RF models in combination with correlation network analysis were used to explore biological processes of healthy aging. The models highlighted established metabolites, like steroids, amino acids, and free fatty acids as well as novel metabolites and pathways. Finally, we show that metabolic profiles of twins become more dissimilar with age which provides insights into nongenetic age‐related variability in metabolic profiles in response to environmental exposure.     

A new application of isoindole fluorophore derivative in sitagliptin antidiabetic medication assay: Application to dosage forms and biological fluid evaluation

Dipeptidyl peptidase-4 enzyme suppressant is a unique category of oral antidiabetic medication. Sitagliptin (STG) is a perfect member of this category and is pharmaceutically marketed alone or in combination with metformin. Here, the ideal application of an isoindole derivative for STG assay was developed using a feasible, easy-to-use, economic, and affordable method. STG as an amino group donor can form a luminescent derivative: isoindole on interaction with o-phthalaldehyde and the existence of (2-mercaptoethanol) 0.02% (v/v) as a thiol group donor. Excitation (339.7 nm) and emission (434.6 nm) wavelengths were used to monitor the isoindole fluorophore yield; moreover, each experimental variable was carefully investigated and adjusted. The calibration graph was constructed by plotting fluorescence intensities against STG concentrations, and controlled linearity was observed at concentrations ranging from 50 to 1000 ng/ml. The International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use guidelines were analyzed in depth to prove the technique validation. The implementation of the present technique was extended successfully to the evaluation of various types of STG dose forms and spiking samples of human plasma and urine. The developed technique was shown to be an effective, simple, and quick replacement for quality control and clinical study evaluation of STG.