Biomedical subjects

Xinghua Lu

Publications and source records attributed to Xinghua Lu.

18 recordsLinked to original sources

PubMed2026

Knowledge-guided contextual gene set analysis with large language models.

MOTIVATION: Gene set analysis (GSA) is a foundational approach for interpreting genomic data of diseases by linking genes to biological processes. However, conventional GSA methods overlook clinical context of the analyses, often generating long lists of enriched pathways with redundant, nonspecific, or irrelevant results. Interpreting these requires extensive, ad-hoc manual effort, reducing both reliability and reproducibility. RESULTS: We introduce cGSA, a novel AI-driven framework that enhances GSA by incorporating context-aware pathway prioritization. cGSA integrates gene cluster detection, enrichment analysis, and large language models to identify pathways that are not only statistically significant but also biologically meaningful. Benchmarking on 102 curated gene sets across 19 diseases and ten disease-related biological mechanisms shows that cGSA outperforms baseline methods by over 30%, with expert validation confirming its increased precision and interpretability. Two independent case studies in melanoma and breast cancer further demonstrate its potential to uncover context-specific insights and support targeted hypothesis. AVAILABILITY AND IMPLEMENTATION: The demo website is publicly available at https://www.ncbi.nlm.nih.gov/CBBresearch/Lu/Demo/cGSA/, while the data and code can be accessed at https://github.com/ncbi-nlp/cGSA.

Large Language Models↗

PubMed2025

Knowledge-guided Contextual Gene Set Analysis Using Large Language Models.

Gene set analysis (GSA) is a foundational approach for interpreting genomic data of diseases by linking genes to biological processes. However, conventional GSA methods overlook clinical context of the analyses, often generating long lists of enriched pathways with redundant, nonspecific, or irrelevant results. Interpreting these requires extensive, ad-hoc manual effort, reducing both reliability and reproducibility. To address this limitation, we introduce cGSA, a novel AI-driven framework that enhances GSA by incorporating context-aware pathway prioritization. cGSA integrates gene cluster detection, enrichment analysis, and large language models to identify pathways that are not only statistically significant but also biologically meaningful. Benchmarking on 102 manually curated gene sets across 19 diseases and ten disease-related biological mechanisms shows that cGSA outperforms baseline methods by over 30%, with expert validation confirming its increased precision and interpretability. Two independent case studies in melanoma and breast cancer further demonstrate its potential to uncover context-specific insights and support targeted hypothesis generation.

Journal Article↗

PubMed2007

Accurate discrimination of pancreatic ductal adenocarcinoma and chronic pancreatitis using multimarker expression data and samples obtained by minimally invasive fine needle aspiration.

To augment cytological diagnosis of pancreatic ductal adenocarcinoma (PDAC) in tissue samples obtained by minimally invasive endoscopic ultrasound-guided fine needle aspiration, we investigated whether a small set of molecular markers could accurately distinguish PDAC from chronic pancreatitis (CP). Expression levels of 29 genes were first determined by quantitative real-time RT-PCR in a training set of tissues in which the final diagnosis was PDAC (n=20) or CP (n=10). Using receiver operator characteristic curve analysis, we determined that the single gene with the highest diagnostic accuracy for discrimination of CP vs. PDAC in the training study was urokinase plasminogen activator receptor (UPAR; AUC value = 0.895, 95% CI=0.728-0.976). In the set of test tissues (n=14), the accuracy of UPAR decreased to 79%. However, we observed that the addition of 6 genes (EPCAM2, MAL2, CEA5, CEA6, MSLN and TRIM29; referred to as the 6-gene classifier) to UPAR resulted in high accuracy in both training and testing sets. Excluding 3 samples (out of 44; 7%) for which results of the UPAR/6-gene classifier were "undefined," the accuracy of the UPAR/6-gene classifier was 100% in training samples (n=29), 92% in 12 test samples (p=0.004 that results were randomly generated; p=0.046 that the UPAR/6-gene classifier was comparable to UPAR alone; chi2 test), 100% in 3 samples for which the initial cytological diagnosis was "suspicious" and 98% (40/41) overall. Our results provide evidence that molecular marker expression data can be used to augment cytological analysis.

Adult↗

PubMed2006

Enhancing text categorization with semantic-enriched representation and training data augmentation.

OBJECTIVE: Acquiring and representing biomedical knowledge is an increasingly important component of contemporary bioinformatics. A critical step of the process is to identify and retrieve relevant documents among the vast volume of modern biomedical literature efficiently. In the real world, many information retrieval tasks are difficult because of high data dimensionality and the lack of annotated examples to train a retrieval algorithm. Under such a scenario, the performance of information retrieval algorithms is often unsatisfactory, therefore improvements are needed. DESIGN: We studied two approaches that enhance the text categorization performance on sparse and high data dimensionality: (1) semantic-preserving dimension reduction by representing text with semantic-enriched features; and (2) augmenting training data with semi-supervised learning. A probabilistic topic model was applied to extract major semantic topics from a corpus of text of interest. The representation of documents was projected from the high-dimensional vocabulary space onto a semantic topic space with reduced dimensionality. A semi-supervised learning algorithm based on graph theory was applied to identify potential positive training cases, which were further used to augment training data. The effects of data transformation and augmentation on text categorization by support vector machine (SVM) were evaluated. RESULTS AND CONCLUSION: Semantic-enriched data transformation and the pseudo-positive-cases augmented training data enhance the efficiency and performance of text categorization by SVM.

Algorithms↗

PubMed2006

A multivariate prediction model for microarray cross-hybridization.

BACKGROUND: Expression microarray analysis is one of the most popular molecular diagnostic techniques in the post-genomic era. However, this technique faces the fundamental problem of potential cross-hybridization. This is a pervasive problem for both oligonucleotide and cDNA microarrays; it is considered particularly problematic for the latter. No comprehensive multivariate predictive modeling has been performed to understand how multiple variables contribute to (cross-) hybridization. RESULTS: We propose a systematic search strategy using multiple multivariate models [multiple linear regressions, regression trees, and artificial neural network analyses (ANNs)] to select an effective set of predictors for hybridization. We validate this approach on a set of DNA microarrays with cytochrome p450 family genes. The performance of our multiple multivariate models is compared with that of a recently proposed third-order polynomial regression method that uses percent identity as the sole predictor. All multivariate models agree that the 'most contiguous base pairs between probe and target sequences,' rather than percent identity, is the best univariate predictor. The predictive power is improved by inclusion of additional nonlinear effects, in particular target GC content, when regression trees or ANNs are used. CONCLUSION: A systematic multivariate approach is provided to assess the importance of multiple sequence features for hybridization and of relationships among these features. This approach can easily be applied to larger datasets. This will allow future developments of generalized hybridization models that will be able to correct for false-positive cross-hybridization signals in expression experiments.

Algorithms↗

PubMed2006

Identifying biological concepts from a protein-related corpus with a probabilistic topic model.

BACKGROUND: Biomedical literature, e.g., MEDLINE, contains a wealth of knowledge regarding functions of proteins. Major recurring biological concepts within such text corpora represent the domains of this body of knowledge. The goal of this research is to identify the major biological topics/concepts from a corpus of protein-related MEDLINE titles and abstracts by applying a probabilistic topic model. RESULTS: The latent Dirichlet allocation (LDA) model was applied to the corpus. Based on the Bayesian model selection, 300 major topics were extracted from the corpus. The majority of identified topics/concepts was found to be semantically coherent and most represented biological objects or concepts. The identified topics/concepts were further mapped to the controlled vocabulary of the Gene Ontology (GO) terms based on mutual information. CONCLUSION: The major and recurring biological concepts within a collection of MEDLINE documents can be extracted by the LDA model. The identified topics/concepts provide parsimonious and semantically-enriched representation of the texts in a semantic space with reduced dimensionality and can be used to index text.

Abstracting and Indexing↗

PubMed2006

Distinct roles for de novo versus hydrolytic pathways of sphingolipid biosynthesis in Saccharomyces cerevisiae.

Saccharomyces cerevisiae produces the sphingolipid ceramide by de novo synthesis as well as by hydrolysis of complex sphingolipids by Isc1p (inositolphosphoceramide-phospholipase C), which is homologous with the mammalian neutral sphingomyelinases. Though the roles of sphingolipids in yeast stress responses are well characterized, it has been unclear whether Isc1p contributes to stress-induced sphingolipids. The present study was undertaken in order to distinguish the relative roles of de novo sphingolipid biosynthesis versus Isc1p-mediated sphingolipid production in the heat-stress response. Ceramide production was measured at normal and increased temperature in an ISC1 deletion and its parental strain (ISC1 being the gene that codes for Isc1p). The results showed that Isc1p contributes specifically to the formation of the C24-, C24:1- and C26-dihydroceramide species. The interaction between these two pathways of sphingolipid production was confirmed by the finding that ISC1 deletion is synthetically lethal with the lcb1-100 mutation. Interestingly, Isc1p did not contribute significantly to transient cell-cycle arrest or growth at elevated temperature, responses known to be regulated by the de novo pathway. In order to define specific contributions of ISC1, microarray hybridizations were performed, and analyses showed misregulation of genes involved in carbon source utilization and sexual reproduction, which was corroborated by defining a sporulation defect of the isc1Delta strain. These results indicate that the two pathways of ceramide production in yeast interact, but differ in their regulation of ceramides of distinct molecular species and serve distinct cellular functions.

Gene Deletion↗

PubMed2005

Spatially dependent inelastic tunneling in a single metallofullerene.

We have measured the elastic and inelastic tunneling properties of isolated Gd@C(82) molecules on Ag(001) using cryogenic scanning tunneling spectroscopy. We find that the dominant inelastic channel is spatially well localized to a particular region of the molecule. Ab initio pseudopotential density-functional theory calculations indicate that this channel arises from a vibrational cage mode. We further show that the observed inelastic tunneling localization is explained by strong localization in the molecular electron-phonon coupling to this mode.

Journal Article↗

PubMed2005

Genetic polymorphisms in methylenetetrahydrofolate reductase and thymidylate synthase and risk of pancreatic cancer.

BACKGROUND & AIMS: Human pancreatic cancer might be associated with folate deficiency and impaired metabolism. We tested this hypothesis by examining the contribution of functional polymorphisms in methylenetetrahydrofolate reductase (MTHFR) and thymidylate synthase (TS) to risk of cancer. METHODS: DNA from 163 pancreatic cancer patients and 337 control subjects was genotyped for MTHFR (677C > T and 1298A > C) and TS (5'-untranslated region tandem repeat and G/C). Association with risk of pancreatic cancer was estimated by logistic regression. All statistical tests were two-sided. RESULTS: We observed an increased risk of pancreatic cancer associated with the MTHFR 677CT (odds ratio [OR], 2.60; 95% confidence interval [CI], 1.61-4.29; P = .0005) or 677TT (OR, 5.12; 95% CI, 2.94-9.10; P < .0001) genotype compared with the MTHFR CC genotype. An increased risk of pancreatic cancer was also associated with the TS 3Rc/3Rc genotype (OR, 2.19, 95% CI, 1.13-4.31; P = .022) compared with the TS 3Rg/3Rg genotype. Joint effect between MTHFR C677T polymorphism and smoking or drinking increased risk of pancreatic cancer in a super-multiplicative manner. The ORs for smoking, the polymorphism, and both factors combined were 0.70 (95% CI, 0.30-1.63), 2.17 (95% CI, 1.17-4.21), and 3.10 (95% CI, 1.54-6.51), respectively. This joint effect was much stronger in heavy smokers (OR, 6.69; 95% CI, 3.39-13.63; P < .0001). The ORs for drinking, the polymorphism, and both factors combined were 0.98 (95% CI, 0.40-2.30), 2.81 (95% CI, 1.65-4.98), and 4.39 (95% CI, 2.25-8.78), respectively. CONCLUSION: The MTHFR and TS polymorphisms are genetic determinants for developing pancreatic cancer.

Adult↗

PubMed2005

[Physiological characteristics of nitrogen nutrition and stress-resistance of film-mulched rice in various ecological regions of Zhejiang Province].

The study showed that different ecological environment and cultivation system in various ecological regions of Zhejiang Province resulted in some different physiological characteristics of nitrogen nutrition and stress-resistance, especially in the aspect of NO3(-)-N and NH4+-N concentrations, between film-mulched and conventional flooded rice. Owing to the heat stress in Hangjiahu plain, the NO3(-)-N concentration of film-mulched rice decreased to some extent, but NH4+-N concentration increased markedly at tillering, jointing and booting stages, compared to conventional flooded rice. In Jinqu basin, the NO3(-)-N concentration of film-mulched rice at booting stage was higher, while the NH4+-N concentration in its roots was notably lower than those of conventional flooded rice, with NH4+-N concentration in its basal stems and leaves somewhat increased. Generally, the glutamine synthetase (GS) and nitrate reductase (NR) activities in film-mulched rice leaves were enhanced at booting stage, while malondiadehyde (MDA), soluble sugar (SS) and proline (Pro) concentrations had little changes. In conclusion, film-mulched cultivation was beneficial to the rice growth and its high yielding.

Biomass↗

PubMed2004

Automatic annotation of protein motif function with Gene Ontology terms.

BACKGROUND: Conserved protein sequence motifs are short stretches of amino acid sequence patterns that potentially encode the function of proteins. Several sequence pattern searching algorithms and programs exist foridentifying candidate protein motifs at the whole genome level. However, a much needed and important task is to determine the functions of the newly identified protein motifs. The Gene Ontology (GO) project is an endeavor to annotate the function of genes or protein sequences with terms from a dynamic, controlled vocabulary and these annotations serve well as a knowledge base. RESULTS: This paper presents methods to mine the GO knowledge base and use the association between the GO terms assigned to a sequence and the motifs matched by the same sequence as evidence for predicting the functions of novel protein motifs automatically. The task of assigning GO terms to protein motifs is viewed as both a binary classification and information retrieval problem, where PROSITE motifs are used as samples for mode training and functional prediction. The mutual information of a motif and aGO term association is found to be a very useful feature. We take advantage of the known motifs to train a logistic regression classifier, which allows us to combine mutual information with other frequency-based features and obtain a probability of correct association. The trained logistic regression model has intuitively meaningful and logically plausible parameter values, and performs very well empirically according to our evaluation criteria. CONCLUSIONS: In this research, different methods for automatic annotation of protein motifs have been investigated. Empirical result demonstrated that the methods have a great potential for detecting and augmenting information about the functions of newly discovered candidate protein motifs.

Amino Acid Motifs↗

PubMed2004

A study of statistical methods for function prediction of protein motifs.

Automatic discovery of new protein motifs (i.e. amino acid patterns) is one of the major challenges in bioinformatics. Several algorithms have been proposed that can extract statistically significant motif patterns from any set of protein sequences. With these methods, one can generate a large set of candidate motifs that may be biologically meaningful. This article examines methods to predict the functions of these candidate motifs. We use several statistical methods: a popularity method, a mutual information method and probabilistic translation models. These methods capture, from different perspectives, the correlations between the matched motifs of a protein and its assigned Gene Ontology terms that characterise the function of the protein. We evaluate these different methods using the known motifs in the InterPro database. Each method is used to rank candidate terms for each motif. We then use the expected mean reciprocal rank to evaluate the performance. The results show that, in general, all these methods perform well, suggesting that they can all be useful for predicting the function of an unknown motif. Among the methods tested, a probabilistic translation model with a popularity prior performs the best.

Algorithms↗

PubMed2003

Spatially mapping the spectral density of a single C60 molecule.

We have used scanning tunneling spectroscopy to spatially map the energy-resolved local density of states of individual C60 molecules on the Ag(100) surface. Spectral maps were obtained for molecular states derived from the C60 HOMO, LUMO, and LUMO+1 orbitals, revealing new details of the spatially inhomogeneous C60 local electronic structure. Spatial inhomogeneities are explained using ab initio pseudopotential density functional calculations. These calculations emphasize the need for explicitly including the C60-Ag interaction and STM tip trajectory to understand the observed C60 local electronic structure.

Journal Article↗

PubMed2003

Unexpected mobility variation among individual secretory vesicles produces an apparent refractory neuropeptide pool.

Most stored neuropeptide cannot be released from nerve terminals suggesting the existence of a refractory pool of dense core vesicles (DCVs). Past fluorescence photobleaching recovery, single particle tracking and release experiments suggested that the refractory neuropeptide pool corresponds to a distinct immobile fraction of cytoplasmic DCVs. However, tracking of hundreds of individual green fluorescent protein-labeled neuropeptidergic vesicles by wide-field or evanescent-wave microscopy shows that a separate immobile fraction is not evident. Instead, the DCV diffusion coefficient (D) distribution is unusually broad and asymmetric. Furthermore, the distribution shifts with a release facilitator. This unexpected variation, which could reflect heterogeneity among vesicles or in their medium, is shown to generate the appearance of a regulated refractory neuropeptide pool.

Animals↗

PubMed2003

Calcium requirements for exocytosis do not delimit the releasable neuropeptide pool.

Recently, it was proposed that secretory vesicles have widely varying Ca(2+) thresholds for exocytosis. This model can explain adaptation of secretory responses and predicts that incomplete release is a consequence of insufficient Ca(2+). However, membrane capacitance-based measurements have not supported varying Ca(2+) thresholds. Here, Green Fluorescent Protein (GFP) imaging is used to test whether a Ca(2+) limitation determines the size of the releasable neuropeptide pool in differentiated PC12 cells. We show that depolarization-evoked release correlates with failure to sustain fully elevated [Ca(2+)](i). However, this is coincidental because release remains incomplete when [Ca(2+)](i) is maintained at a relatively high level by application of an ionophore or by dialysis with a buffered Ca(2+) solution. Furthermore, in contradiction with the existence of high threshold vesicles, stimulating maximal release with moderate [Ca(2+)](i) prevents secretory responses to large increases in [Ca(2+)](i) induced by photolysis of the caged dimethoxynitrophenyl-EGTA-4 (DMNPE-4). Thus, optical measurements show that limited capacity for neuropeptide release in response to depolarization is not caused by an insufficient duration of [Ca(2+)](i) elevation or by variation among vesicles in Ca(2+) sensitivity for exocytosis.

Action Potentials↗

PubMed2002

Physical mobilization of secretory vesicles facilitates neuropeptide release by nerve growth factor-differentiated PC12 cells.

It has been speculated that neurosecretion can be enhanced by increasing the motion, and hence, the availability of cytoplasmic secretory vesicles. However, facilitator-induced physical mobilization of secretory vesicles has not been observed directly in living cells, and recent experimental results call this hypothesis into question. Here, high resolution green fluorescent protein (GFP)-based measurements in nerve growth factor-differentiated PC12 cells are used to test whether altering dense core vesicle (DCV) motion affects neuropeptide release. Experiments with mycalolide B and jasplakinolide demonstrate that neuropeptidergic DCV motion at the ends of processes is proportional to F-actin. Furthermore, Ba2+ increases DCV mobility without detectably modifying F-actin. Finally, we show that altering DCV motion by changing F-actin or stimulating with Ba2+ proportionally changes sustained neuropeptide release. Therefore, increasing DCV mobility facilitates prolonged neuropeptide release.

Actins↗

PubMed2002

Nerve growth factor-induced differentiation changes the cellular organization of regulated Peptide release by PC12 cells.

PC12 cells, like endocrine chromaffin cells, undergo neuronal-like differentiation in response to nerve growth factor (NGF). Here we report that this phenotype conversion produces major changes in release of a green fluorescent protein-tagged neuropeptide-hormone. First, the spatial distribution of the releasable pool is altered; peptide release from untreated cells is supported predominantly by membrane-proximal vesicles, whereas a diffuse pool at the ends of processes is used by NGF-treated cells. Second, the time course of release evoked by photolysis of caged Ca(2+) is faster after differentiation. High-resolution measurements suggest that a slow step before membrane fusion dominates the kinetics of release in untreated cells. Finally, the effect of actin microfilament depolymerization on total release is altered by NGF treatment. This implies that the mechanism that limits the size of the releasable pool is altered by phenotype conversion. Therefore, the cellular organization of peptide release is plastic and changes in response to NGF. This flexibility may be used to generate cell-specific release properties.

Actin Cytoskeleton↗

PubMed2002

Detecting K-ras and p53 gene mutation from stool and pancreatic juice for diagnosis of early pancreatic cancer.

OBJECTIVE: To explore new methods for the early diagnosis of pancreatic cancer through detection of K-ras and p53 mutations in pancreatic juice and stool. METHODS: 201 patients in PUMC Hospital from 1994 - 2000 and 60 control individuals were enrolled in this study. K-ras point mutation was detected by PCR-RFLP while p53 mutation was detected by PCR-SSCP. RESULTS: K-ras mutation was found in pancreatic juice in 87.8% (36/41) of pancreatic cancer patients and 23.5% (4/17) of benign pancreatic disease patients. In 261 stool specimens, amplification found mutations successfully in 235 patients (90%). K-ras mutation was found in stool in 88% (66/75) of pancreatic cancer patients, 51.1% (24/47) of benign pancreatic disease patients and 19.6% (9/46) of normal individuals. p53 mutation was found in pancreatic juice in 47.4% (18/38) of pancreatic cancer patients and 12.5% (2/16) of benign pancreatic disease patients. p53 mutation was found in stool in 37.1% (23/62) and 19.1% (4/21) of chronic pancreatitis patients. CONCLUSION: K-ras mutation in pancreatic juice has higher diagnosis sensitivity and specificity, and therefore may be used as a supplement in the diagnosis of pancreatic cancer. Detection of K-ras mutation combined with p53 mutation in stool can aid in the screening of pancreatic cancer.

Feces↗