Improving sentence simplification model with ordered neurons network

Sentence simplification is an essential task in natural language processing and aims to simplify complex sentences while retaining their primary meanings.To date,the main research works on sentence simplification models have been based on sequence-to-sequence(Seq2Seq)models.However,these Seq2Seq models are incapable of analysing the hierarchical structure of sentences,which is of great significance for sentence simplification.The problem can be addressed with an ON-MULTI-STAGE model constructed based on the improved MULTI-STAGE encoder model.In this model,an ordered neurons network is introduced and can provide sentence-level structural infor-mation for the encoder and decoder.A weak attention connection method is then employed to make the decoder use the sentence-level structural details.Experimental results on two open data sets demonstrated that the constructed model outperforms the state-of-the-art baseline models in sentence simplification.

Improving word vector model with part-of-speech and dependency grammar information

Part-of-speech(POS)and dependency grammar(DG)are the basic components of natural language processing.However,current word vector models have not made full use of both POS information and DG information,and hence the models’performances are limited to some extent.The authors first put forward the concept of POS vector,and then,based on continuous bag-of-words(CBOW),constructed four models:CBOW+P,CBOW+PW,CBOW+G,and CBOW+G+P to incorporate POS information and DG information into word vectors.The CBOW+P and CBOW+PW models are based on POS tagging,the CBOW+G model is based on DG parsing,and the CBOW+G+P model is based on POS tagging and DG parsing.POS information is integrated into the training process of word vectors through the POS vector to solve the problem of the POS similarity being difficult to measure.The POS vector correlation coefficient and distance weighting function are used to train the POS vector as well as the word vector.DG information is used to correct the information loss caused by fixed context windows.Dependency relations weight is used to measure the difference of dependency relations.Experiments demonstrated the superior performance of their models while the time complexity is still kept the same as the base model of CBOW.

Peptidome data-driven comprehensive individualized monitoring of membranous nephropathy with machine learning

As the most common pathological type of nephrotic syndrome,membranous nephropathy(MN)presents diversity in progression trends,facing severe complications.The precise discrimination of MN from healthy people,other types of nephrotic syndrome or those with therapeutic remission has always been huge challenge in clinics,not to mention comprehensive individualized monitoring relied on minimally invasive molecular detection means.Herein,we construct a functionalized pore architecture to couple with machine learning to aid all-round peptidome enrichment and data profiling from hundreds of human serum samples,and finally establish a set of defined peptide panel consisting of 12 specific feature signals.In addition to the realization of above-mentioned precise discrimination with more than 97%of sensitivity,88%of accuracy and f1 score,the simultaneously comprehensive individualized monitoring for MN can also be achieved,including conventionally screening diagnosis,congeneric distinction and prognostic evaluation.This work greatly advances the development of peptidome data-driven individualized monitoring means for complex diseases and undoubtedly inspire more devotion into molecular detection field.

Removal of Microcystin-LR in lake water sample by hydrophilic mesoporous silica composites under high-throughput MALDI-TOF MS detection platform

Microcystins(MCs),a family of cyclic heptapeptide cyanotoxins,exists in aquatic environment where cyanobacterial bloom happens,which will accumulate in aquatic organisms and transfer through the food chain to higher trophic levels,posing a health risk to both animals and human bodies.Among various MCs,Microcystin-LR(MC-LR)is worthiest studied for its strong toxicity,ubiquity and widespread.Here in this work,iminodiacetic acid(IDA)decorated magnetic mesoporous silica(mSiO_(2))nanocomposites(Fe_(3)O_(4)@mSiO_(2)-IDA)were facilely synthesized which possessed the merits of large surface area(188.21 m^(2)/g),accessible porosity(2.66 nm),excellent hydrophilicity and rapid responsiveness to magnetic field.Then the composites were successfully employed to the removal process of Microcystin-LR in real water samples followed by Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry(MALDI-TOF MS)analysis,achieving the removal efficiency above 92.5%even after ten recycles of the composites.It provided a potential method for removing MC-LR in aqueous environment with high effectiveness,lower costs and less secondary contamination.

Probing region-resolved heterogeneity of phosphoproteome in human lens by hybrid metal organic frameworks

Phosphorylation plays crucial parts in lenticular biological function.Getting knowledge of region-resolved phosphoproteome contributes to better comprehending the pathogenesis.Here,we prepared the hybrid metal organic frameworks(HMOFs)for probing the region-resolved heterogeneity of phosphoproteome in human lens.1334 phosphosites corresponding to 564 phosphoproteins,1160 phosphosites corresponding to 316 phosphoproteins and 517 phosphosites corresponding to 205 phosphoproteins were identified in capsule,cortex and nucleus,respectively,providing the relatively extensive distribution mapping of phosphorylation in human lens for the first time.The label-free quantification experiments and principal component analysis presented differential expression of phopshoproteins in three subregions.For instance,α-crystallin,β-crystallin and fibrillin-1 closely associated with cataract and Marfan syndrome showed disparate spatial distribution.The preferential phosphoproteins in capsule,cortex and nucleus were involved in cytoskeleton organization,metabolic process and lens development in camera-type eye,respectively.This work first provided a general overview of region-resolved phosphoproteome of human lens.

Diagnosis and subtype classification on serum peptide fingerprints by mesoporous polydopamine with built-in metal-organic framework

Crohn's disease(CD)as a big issue to public health needs an accurate diagnosis urgently that is the common challenge among internal diseases.Herein,we design a mesoporous polydopamine with builtin metal-organic frameworks(dubbed MMP-b-MOFs)to combine with high-throughput mass spectrometry to extract serum peptide fingerprints from CD and healthy controls(HC).Benefitting by the sizeexclusion and strong hydrophilicity of MMP-b-MOFs,the extracted peptide fingerprints present extremely high quality.CD and HC are explicitly discriminated with orthogonal partial least squares discriminant analysis(OPLS-DA),the corresponding area under the curve(AUC)value is 1.000.Moreover,eight peptides with clear identity are screened out and achieve the accurate diagnosis and subtype classification of CD,with all AUC values up to 1.000.Moreover,the unsupervised model is also established to precisely classify HC and CD based on these eight clearly identified peptides.This work brings great benefits for clinical detection especially internal medicine.

Specific enrichment of urinary exosomes and exosomal glycopeptides by coefficient affinity of integrated L-cysteine and titania

Exosome and inclusive cargoes have manifested significant function in different biological events. In particular, glycopeptides in exosome are closely associated with occurrence and development of various diseases. Developing advanced tools is highly desired to enrich glycopeptides from exosomes, and enrich exosomes from complex biological samples as well. In this work, integration of L-cysteine and titania onto the surface of magnetic nanoparticles is designed to realize the coefficient affinity towards exosomes and inclusive glycopeptides. Benefiting from the synergistic affinity, we separate exosomes from human urine concentrate directly, which was proved by the detection of three typical antigen markers of exosomes. Furthermore, hardly any exosomes remained on materials after ultrasonication, which confirmed the good capture performance of Fe_(3)O_(4)@TiO_(2)@L-Cys and high release effect of direct lysis.Moreover, 146 glycopeptides corresponding to 77 glycoproteins were successfully identified from captured exosomes. These satisfactory results will inspire more efforts to be devoted to this field and will be extremely helpful to in-depth information excavation of biological markers, especially disease-related ones, through exosomes and inclusive glycopeptides.

Sulfonic acid-based metal organic framework functionalized magnetic nanocomposite combined with gas chromatography-electron capture detector for extraction and determination of organochlorine

The metal organic framework functionalized with sulfonic acid was combined with magnetic nanoparticles to fabricate a new nanocomposite(denoted as Fe3O4@PDA@Zr-SO3H).By combining with gas chromatography-electron capture detector,the resulting Fe3O4@PDA@Zr-SO3H nanocomposite was successfully used as a high-efficiency adsorbent for pre-concentrating eight organochlorine pesticides from water sample in environment.Apart from the ability of fast separation,the as-prepared Fe3O4@PDA@Zr-SO3H nanocomposite also exhibited high adsorption capacity for organochlorine pesticides.With the use of optimal experimental conditions,the linear relationship can be obtained in the range of 0.05~300μg/L,the correlation coefficient was over 0.9978,and the relative standard deviation was located in 2.5%-7.7%.Moreover,the limit of detection and quantification was between0.005-0.016μg/L and 0.017~0.050μg/L.Finally,the nanocomposite was used for the determination of organochlorine pesticides from environmental water samples,and displayed the recovery of 82%-118%.

Inherently hydrophilic mesoporous channel coupled with metal oxide for fishing endogenous salivary glycopeptides and phosphopeptides

Both glycosylation and phosphorylation exert crucial rule in multitudinous biological processes.For in-depth profiling of glycosylation and phosphorylation,a magnetic metal oxide is effectively coupled with inherently hydrophilic mesoporous channels(denoted as Fe_(3)O_(4)@TiO_(2)@mSiO_(2)-TSG).Based on the mechanism of hydrophilic interaction liquid chromatography(HILIC)and metal oxide affinity chromatography(MOAC),the Fe_(3)O_(4)@TiO_(2)@mSiO_(2)-TSG nanomaterial shows high capacity for simultaneously enriching glycopeptides and phosphopeptides.With human saliva collected in successive four days as practical biological sample,endogenous glycopeptides and phosphopeptides are efficiently enriched.Further gene ontology analysis reveals that the identified endogenous glycopeptides and phosphopeptides participate in diverse molecular functions and biological processes.This strategy is anticipated to promote variation analysis of salivary post-translational modifications.

Advances in aptamer-based nanomaterials for separation and analysis of non-genetic biomarkers in biofluids

Body fluids contain a large number of disease markers including proteins,exosomes and cells,the concentration of which indicates the health status of a body system and can be hence applied to predict the degree of seriousness of a disease.However,the content of disease markers is relatively low.For solving this problem,a variety of pretreatment methods have been developed,among which aptamer-based nanomaterials hold great potential for biomarker capture because of various natural properties of nano-substrates such as high specific surface area,as well as the high affinity of aptamer towards target and the higher stability than antibody.In this review,we summarized the recent aptamer-based nanomaterials for the separation and analysis of different biomarkers including proteins,exosomes and cells.More specifically,the standard of classification description was based on the applied substrates including magnetic substrates and non-magnetic substrates.Moreover,the perspective of aptamer-based nanomaterials was discussed.

Metal organic frameworks as advanced extraction adsorbents for separation and analysis in proteomics and environmental research

Human health is always under global spotlight, but now it suffers from severe environmental issue and various diseases.Developing highly selective and effective extraction technique for environmental pollutant removal and target protein/peptide molecules conduces to the solution of environmental contamination and early disease detection, which will promise human health to a great degree. Metal-organic frameworks(MOFs) are typical porous nanomaterials featuring rigid structure, large specific surface area, thermal and chemical stability, numerous active sites and easy functionalization. Moreover, multitudinous combination of metal centers and organic ligands often endows MOFs with structure and function diversity. Owing to such splendid merits, MOFs are able to act as advanced adsorbents for sample-pretreatment in proteomics and environmental field,which indicates great efficiency, reproducibility, convenience and cost-effectiveness. In this review, the novel design of MOFs and their applications in proteomics and environmental research are roughly summarized. We first introduce the representative researches of functional MOFs in protein post-translational modifications including glycosylation and phosphorylation, mainly focusing on hydrophilic interaction liquid chromatography(HILIC) or metal-oxide affinity chromatography(MOAC). The important role of MOFs in enzyme immobilization is also covered. Then MOF-based adsorptive removal of diverse environmental contaminants embracing plasticizers, pesticides and varied water pollutants(pharmaceutical and personal care products,organic dyes and heavy metals), in which multiple chemical and non-chemical interactions play a vital role, is emphatically discussed. Finally, we analyze the existing problems and challenges in the way of wider practical applications and put forward prospective anticipation.

Metabolic Molecular Diagnosis of Inflammatory Bowel Disease by Synergistical Promotion of Layered Titania Nanosheets with Graphitized Carbon

Due to inefficient diagnostic methods,inflammatory bowel disease(IBD)normally progresses into severe complications including cancer.Highly efficient extraction and identification of metabolic fingerprints are of significance for disease surveillance.In this work,we synthesized a layered titania nanosheet doped with graphitized carbon(2D-GC-mTNS)through a simple one-step assembly process for assisting laser desorption ionization mass spectrometry(LDI-MS)for metabolite analysis.Based on the synergistic effect of graphitized carbon and mesoporous titania,2D-GC-mTNS exhibits good extraction ability including high sensitivity(<1 fmolμL−1)and great repeatability toward metabolites.A total of 996 fingerprint spectra were collected from hundreds of native urine samples(including IBD patients and healthy controls),each of which contained 1220 m/z metabolite features.Diagnostic model was further established for precise discrimination of patients from healthy controls,with high area under the curve value of 0.972 and 0.981 toward discovery cohort and validation cohort,respectively.The 2D-GC-mTNS promotes LDI-MS to be close to clinical application,with rapid speed,minimum sample consumption and free of sample pretreatment.