Biology subjects

Tian, P.

Publications and source records attributed to Tian, P..

At least 19 recordsLinked to original sources

bioRxiv2026

Characterization of cross-family conserved fragments

In a companion study, we introduced ProDive and used it to identify 318,289 high-confidence cross-family segment correspondences across all 25,545 Pfam families. Here we characterize the pervasiveness and biophysical context of the resulting mapped fragment resource. Graph-based community analysis shows that segment reuse is broadly distributed across the protein family universe, with thousands of mostly small modules rather than concentration in a small number of family pairs. Comparison between Pfam families and de novo designed proteins reveals an approximately 2.1-fold enrichment of cross-family segment correspondences relative to the Pfam-Pfam background. Mapped fragments are strongly enriched in helix-dominant secondary structures with intermediate solvent exposure, occupy a narrowed sequence-context entropy regime, and show a significant shift toward lower fragment-associated contact order relative to random control windows, directly supporting a folding-compatible local structural regime. Matched annotation analyses show depletion from known domain-based interaction interfaces and mild depletion from strict DisProt disorder. Direct Start2Fold/HDX and{varphi} -value comparisons are limited in scale but connect the same fragment class to folding-related protection: stability annotations support local structural stabilization that may facilitate folding, whereas folding annotations and transition-state enrichments provide more specific evidence for folding initiation or pathway involvement. Together, these results support a physical-constraint interpretation of local structural conservation, in which reusable compact fragments contribute prominently to local stability and folding-relevant local structure, with folding initiation supported in specific subsets rather than as an exclusive mechanism.

bioinformatics↗

bioRxiv2026

De novo design of flexible protein interactions with GuideFlip

De novo design of protein binders requires a target structure. However, for flexible targets, such as intrinsically disordered proteins, this structure does not exist until the binder has stabilized the interaction. Such targets are therefore difficult for methods that separate structure generation from sequence design. We introduce GuideFlip, which co-designs structure and sequence through guided discrete flow matching: binder residues are assigned progressively while the complex is re-predicted at each step, allowing the evolving interface to affect the design process. GuideFlip reduces the hydrophobic bias of direct AlphaFold optimization and improves in silico success rates over existing approaches. We release a database of binder candidates for 177 human disordered proteins. Experimentally, we obtain de novo binders to the C-terminus of -synuclein and the disordered amino terminus of RBX1 with hit rates of 13.5% and 41.7%, respectively, and we confirm the epitopes of selected binders by NMR and mutagenesis. Applying GuideFlip to flexibility on the binder side, we design a nanobody that binds the agonist-bound {beta}1-adrenergic receptor in the active state, but not the receptor in its inactive state, with a 75% hit rate and cryo-EM structure confirming the design. GuideFlip enables protein design where bound structures emerge only upon binding.

biophysics↗

bioRxiv2026

Mapping bacterial cutinase sequence space by high-throughput screening reveals that PET hydrolysis is a rare property

Poly(ethylene terephthalate) (PET) is one of the most widely produced plastics, and enzymatic depolymerization offers a promising route to closed-loop recycling under mild conditions. However, most known bacterial PET hydrolases belong to a conserved canonical-fold cutinase family, leaving much of alpha/beta-hydrolase diversity unexplored. Here, we mapped bacterial cutinase sequence space by combining bioinformatics-guided sequence selection with high-throughput secretion screening in Bacillus subtilis. A library of 1,120 genes encoding 954 unique bacterial cutinases, spanning canonical- and minimal-fold families, was screened for activity on Impranil DLN and semicrystalline PET. We identified 156 secreted cutinases with polyester activity, broadly distributed across sequence space, but only ten showed detectable PET hydrolysis, all from the canonical-fold family. These PET hydrolases were active at 40-50{degrees}C, preferred alkaline pH, and showed moderate thermostability. Our results demonstrate that PET activity is rare among bacterial cutinases and provide a scalable workflow for discovering diverse enzyme starting points.

biochemistry↗

bioRxiv2026

ProDive reveals pervasive cross-family protein fragment reuse

Cross-family reuse of short protein fragments has been a long-standing mystery whose resolution first demands an algorithm for their systematic detection. Here we introduce ProDive, a closed-form symmetric KL divergence between profile HMMs that enables GPU-accelerated, fragment-level screening across all 25,545 Pfam families. ProDive identifies [~]318,000 cross-family fragment correspondences involving compact cores of 8-13 residues with RMSD values far below random background. Their organisation into diverse graph communities and four-fold enrichment in de novo designed proteins point away from family-specific functions and toward a general biophysical property. Their helix dominance and moderate solvent exposure suggest a role in folding initiation--a link corroborated by overlap with experimentally measured{phi} -values and by a monotonic density gradient across disordered regions. Together, these observations converge on a single explanation: cross-family fragment reuse likely reflects shared requirements for early structure formation during folding, the one biophysical constraint common to all proteins.

bioinformatics↗

bioRxiv2025

Host EV derived lipids promote malaria parasite development via mosquito fatty acid oxidation

Malaria parasites must transition between mammals and mosquitoes, but how they adapt to such distinct environments is unclear. We show that blood from infected mammals buffers this environmental shift by packaging parasite-induced signals into extracellular vesicles (EVs). EVs internalized by mosquito midgut increases expression of VLCAD and raises levels of a 22-carbon fatty acid, driving fatty acid breakdown and acetyl-CoA production. The parasite exploits this metabolic shift to synthesize its own lipids during oocyst growth. Stable isotope tracing confirmed that13C4-labelled behenic acid is degraded and incorporated into parasite fatty acids. Lipidomic profiling further revealed that ceramides enriched in the EVs are key mediators. These findings identify a mechanism by which host bood-derived EVs reprogram mosquito lipid metabolism to promote parasite transmission.

microbiology↗

bioRxiv2025

CoxMDS: Multiple Data Splitting for High-dimensional Mediation Analysis with Survival Outcomes in Epigenome-wide Studies

Causal mediation analysis investigates whether the effect of an exposure on an outcome operates through intermediate variables known as mediators. Although progress has been made in high-dimensional mediation analysis, current methods do not reliably control the false discovery rate (FDR) in finite samples, especially when mediators are moderately to highly correlated or follow non-Gaussian distributions. These challenges frequently arise in DNA methylation studies. We introduce CoxMDS, a multiple data splitting method that uses Cox proportional hazards models to identify putative causal mediators for survival outcomes. CoxMDS ensures finite-sample FDR control even in the presence of correlated or non-Gaussian mediators. Through simulations, CoxMDS is shown to maintain FDR control and achieve higher statistical power compared with existing approaches. In applications to DNA methylation data with survival outcomes, CoxMDS identified eight CpG sites in The Cancer Genome Atlas (TCGA) that are consistent with the hypothesis that DNA methylation may mediate the effect of smoking on lung cancer survival, and two CpG sites in the Alzheimers Disease Neuroimaging Initiative (ADNI) that are consistent with the hypothesis that DNA methylation may mediate the effect of smoking on time to Alzheimers disease conversion.

bioinformatics↗

bioRxiv2025

Therapeutic splice modulation of COL4A5 reinstates collagen IV assembly in an organoid model of X-linked Alport syndrome

Kidney organoids are an emerging tool for disease modeling, especially genetic diseases. Among them, X-linked Alport syndrome (XLAS) is a hematuric nephropathy affecting the glomerular basement membrane (GBM) secondary to pathogenic variations in the COL4A5 gene encoding the 5 subunit of type IV collagen [5(IV)]. In patients carrying pathogenic variations affecting splicing, the use of antisense oligonucleotides (ASOs) offers immense therapeutic hope. In this study, we develop a framework combining the use of patient-derived cells and kidney organoids to provide evidence of the therapeutic efficacy of ASOs in XLAS patients. Using multiomics analysis, we describe the development of GBM in wild-type and mutated human kidney organoids. We show that GBM maturation is a dynamic process, which requires long organoid culture. Then, using semi-automated quantification of 5(IV) at basement membranes in organoids carrying the splicing variants identified in patients, we demonstrate the efficacy of ASO treatment for 5(IV) restoration. These data contribute to our understanding of the development of GBM and pave the way for a therapeutic screening platform for patients.

genetics↗

bioRxiv2025

Field Survey of Anopheles Mosquito and Their Culturable Gut Bacteria in the Border Regions of Yunnan Province, China

Despite the elimination of domestic malaria transmission within China, the border areas of Yunnan Province continue to face a heightened risk of imported cases due to the presence of diverse Anopheles mosquito species and the widespread burden of malaria in neighboring countries. While the mosquito gut microbiota offers a promising avenue for developing anti-malaria strategies, a meticulous investigation of the various Anopheles species and their associated gut microbiota in Yunnans border areas has yet to be undertaken. To address this gap, field surveys were conducted over the years 2020, 2021, and 2023. These surveys resulted in the collection of a substantial sample comprising 4,221 adult female mosquitoes spanning sixteen species with Anopheles sinensis identified as the predominant species. Furthermore, 741 cultivable strains of gut symbiotic bacteria were isolated and classified into four phyla comprising 109 genera with the composition of dominant bacteria varying annually. This study provides valuable insights into the population dynamics of malaria vectors in the border regions of Yunnan province, China, along with the intestinal bacteria they harbor including several bacterial species which exhibit promising anti-plasmodium effects. By leveraging the antagonistic effects of symbiotic bacteria against Plasmodium, our findings contribute to an ongoing global endeavor to develop innovative Anopheles management and malaria prevention strategies.

microbiology↗

bioRxiv2025

Intratumoral Resident Microbes Directly Participate in Resistance to EGFR-TKI Targeted Drug Therapy through Metabolism in NSCLC

EGFR-TKIs targeted therapy represents a cornerstone in the treatment of non-small cell lung cancer (NSCLC). Nonetheless, a subset of patients inevitably develops acquired resistance to EGFR-TKIs, posing a significant clinical challenge. An increasing number of evidence has documented the presence of microbial communities within tumor microenvironments, but the potential interplay between tumor-resident microbiota and resistance to EGFR-TKIs remains poorly understood. Here, we report that prolonged EGFR-TKIs treatment markedly alters the diversity of tumor-resident microbiota, characterized by a notable enrichment of Bacillus mycoides and Pseudomonas veronii. In vitro experiments revealed that co-culture of gefitinib with either B. mycoides or P. veronii significantly attenuated its tumor-suppressive efficacy, both in vitro and in vivo. Mechanistically, gefitinib underwent extensive biotransformation in the presence of these bacteria, yielding redox derivatives and conjugation products, which collectively diminished its antitumor activity. Specifically, B. mycoides metabolized gefitinib into B.mPM9-C33 H45 Cl F N3 O4, while P. veronii generated P.vPM2-C24 H29 Cl N4 O3, both of which were consistently detected in in vivo and in vitro models. Furthermore, heterologous expression of maltose acetyltransferase from B. mycoides in E. coli BL21 conferred the capacity to metabolize gefitinib, underscoring the enzymatic basis of this metabolic conversion. Collectively, these findings demonstrate that B. mycoides and P. veronii drive tumor resistance to EGFR-TKIs through metabolic inactivation of the drug, unveiling a previously unrecognized role of tumor-resident microbiota in modulating therapeutic responses and providing novel insights into the mechanisms underlying EGFR-TKIs resistance in NSCLC.

microbiology↗

bioRxiv2024

Accurate Multiple Sequence Alignment of Ultramassive Genome Sets

With ever increasing sequencing efficiency, there is a pressing need to tackle a presently intractable task of accurate multiple sequence alignment (MSA) for ultramassive genome sets of millions and beyond. Additionally, efficient graph and probabilistic representations for downstream analysis are in dire lack. In light of these challenges, we develop a set of essentially linearly scalable algorithms, including that for constructing directed acyclic graphs, for training tiled profile hidden Markov models and for conducting alignment on such graphs. The power of these algorithms is demonstrated by both significantly improved accuracy and tremendous acceleration of SAR-CoV-2 MSA by three observed to five projected orders of magnitude for genome set sizes ranging from 40,000 to 4 million when compared with widely utilized MAFFT. Future application to other viral species and extension to more complex genomes will prove this algorithm set as a cornerstone for the coming era of ultramassive genome sets.

bioinformatics↗

bioRxiv2024

Granzyme B-based CAR T cells block metastasis by eliminating circulating tumor cells

Chimeric antigen receptor (CAR) T cells have limited efficacy against solid tumors due to the hostile microenvironment. Circulating tumor cells (CTCs) are essential to metastasis, which is the cause of most of cancer-related death. Here, we generated GrB-CAR T cells targeting membrane-bound HSP70 (mHSP70), a highly tumor-specific antigen detected in numerous cancers. GrB-CAR T cells exhibited potent cytotoxicity against a broad spectrum of cancer cell lines and stem-like cancer cells in vitro and effectively inhibited xenograft tumor growth in vivo. Importantly, GrB-CAR T cells markedly decreased the number of CTCs and, therefore, hindered cancer metastasis in spontaneous metastasis models with uncontrollable primary tumor growth, a scenario commonly encountered in clinical trials of CAR T therapies for solid tumors. Furthermore, despite the 100% homology between human and macaque HSP70 protein, the autotransplantation of macaque T cells expressing human GrB-CAR did not cause any obvious toxic effects. These results not only demonstrate GrB-CAR T cells as a safe and effective tactic with broad-spectrum anticancer activity, but also offer strong experimental evidence and proof-of-concept validation for CAR T cell-mediated metastasis inhibition by targeting CTCs.

immunology↗

bioRxiv2023

scPharm: identifying pharmacological subpopulations of single cells for precision medicine in cancers

Intratumour heterogeneity is a major challenge that limits the effectiveness of anticancer therapies, thus compromising treatment outcomes. Single-cell RNA sequencing (scRNA-seq) technology offers a means to capture gene expression profiles at a single-cell resolution, while drug perturbation experiments yield valuable pharmacological data at the bulk cell level. Here, we introduce "scPharm", a computational framework to integrate large-scale pharmacogenomics profiles with scRNA-seq data, for identifying pharmacological subpopulations within a tumour and prioritizing tailored drugs. scPharm assesses the distribution of the identity genes of single cell (Cell-ID) within drug response-determined gene list, which is accomplished using the normalized enrichment score (NES) obtained from Gene Set Enrichment Analysis (GSEA) as the statistic. One key strength of scPharm is rooted in the robust positive correlation between NES statistics and drug response values at single-cell resolution. scPharm successfully identifies sensitive subpopulations in ER-positive and HER2-positive human breast cancer tissues, discovers dynamic changes in resistant subpopulation of human PC9 cells treated with Erlotinib, and expands its prediction capabilities to a mouse model. By a thoroughly comparative evaluation with other single-cell prediction tools, scPharm presents the superior predictive performance and computational efficiency. Furthermore, scPharm offers a unique feature by predicting combination strategies, gauging compensation effects or booster effects between two drugs through the Set covering method, as well as evaluating drug toxicity on healthy cells within the tumour microenvironment. Together, scPharm provides a novel approach to uncover therapeutic heterogeneity within tumours at single-cell resolution and facilitates precision medicine in cancers.

bioinformatics↗

bioRxiv2023

Epigenetic reader ZMYND11 noncanonical function restricts HNRNPA1-mediated stress granule formation and oncogenic activity

ZMYND11 encodes an epigenetic reader of histone methylation, functioning as a transcriptional corepressor. However, whether and how ZMYND11 contributes to cancer progression and therapy remains unclear. Here we report that ZMYND11 downregulation is prevalent in cancers and profoundly correlates with adverse prostate cancer patient outcomes. Depletion of ZMYND11 promotes tumor cell growth, migration and invasion in vitro as well as tumor formation and metastasis in vivo. Mechanistically, we find that ZMYND11 exhibits tumor suppressive roles through recognizing arginine-194-methylated HNRNPA1 dependent on its MYND domain, thereby squeezing HNRNPA1 in nucleus and inhibiting the formation of stress granules in cytoplasm. Furthermore, ZMYND11 antagonizes HNRNPA1-driven high PKM2/PKM1 ratio and counteracts PKM2-induced aggressive tumor phenotype. Remarkably, ZMYND11 recognition of HNRNPA1 could be disrupted by pharmaceutical inhibition of arginine methyltransferase PRMT5 while ZMYND11 low-expressing tumors are sensitive to the treatment of PRMT5 inhibitors. Collectively, our study unravels a novel and noncanonical function of ZMYND11 as the nonhistone methyl reader and discovers a mechanism for the requirement of arginine-methylation-mediated ZMYND11-HNRNPA1 association to restrict tumor progression and offers cancer therapeutic targets and potential biomarkers.

cancer biology↗

bioRxiv2023

FLOP: Tasks for Fitness Landscapes Of Protein wildtypes

Protein engineering has the potential to create optimized protein variants with improved properties and function. An initial step in the protein optimization process typically consists of a search among natural (wildtype) sequences to find the naturally occurring proteins with the most desirable properties. Promising candidates from this initial discovery phase then form the basis of the second step: a more local optimization procedure, exploring the space of variants separated from this candidate by a number of mutations. While considerable progress has been made on evaluating machine learning methods on single protein datasets, benchmarks of data-driven approaches for global fitness landscape exploration are still lacking. In this paper, we have carefully curated a representative benchmark dataset, which reflects industrially relevant scenarios for the initial wildtype discovery phase of protein engineering. We focus on exploration within a protein family, and investigate the downstream predictive power of various protein representation paradigms, i.e., protein language model-based representations, structure-based representations, and evolution-based representations. Our benchmark highlights the importance of coherent split strategies, and how we can be misled into overly optimistic estimates of the state of the field. The codebase and data can be accessed via https://github.com/petergroth/FLOP.

bioinformatics↗

bioRxiv2023

Assessing the performance of protein regression models

To optimize proteins for particular traits holds great promise for industrial and pharmaceutical purposes. Machine Learning is increasingly applied in this field to predict properties of proteins, thereby guiding the experimental optimization process. A natural question is: How much progress are we making with such predictions, and how important is the choice of regressor and representation? In this paper, we demonstrate that different assessment criteria for regressor performance can lead to dramatically different conclusions, depending on the choice of metric, and how one defines generalization. We highlight the fundamental issues of sample bias in typical regression scenarios and how this can lead to misleading conclusions about regressor performance. Finally, we make the case for the importance of calibrated uncertainty in this domain.

bioinformatics↗

bioRxiv2022

CoxMKF: A Knockoff Filter for High-Dimensional Mediation Analysis with a Survival Outcome in Epigenetic Studies

MotivationIt is of scientific interest to identify DNA methylation CpG sites that might mediate the effect of an environmental exposure on a survival outcome in high-dimensional mediation analysis. However, there is a lack of powerful statistical methods that can provide a guarantee of false discovery rate (FDR) control in finite-sample settings. ResultsIn this article, we propose a novel method called CoxMKF, which applies aggregation of multiple knockoffs to a Cox proportional hazards model for a survival outcome with high-dimensional mediators. The proposed CoxMKF can achieve FDR control even in finite-sample settings, which is particularly advantageous when the sample size is not large. Moreover, our proposed CoxMKF can overcome the randomness of the unstable model-X knockoffs. Our simulation results show that CoxMKF controls FDR well in finite samples. We further apply CoxMKF to a lung cancer data set from The Cancer Genome Atlas (TCGA) project with 754 subjects and 365 306 DNA methylation CpG sites, and identify four DNA methylation CpG sites that might mediate the effect of smoking on the overall survival among lung cancer patients. AvailabilityThe R package CoxMKF is publicly available at https://github.com/MinhaoYaooo/CoxMKF. Contactzl2509@cumc.columbia.edu.

bioinformatics↗

bioRxiv2022

Variable Selection via Grace-AKO with Applications in Genomics

MotivationVariable selection is a common statistical approach to identifying genes associated with clinical outcomes of scientific interest. There are thousands of genes in genomic studies, while only a limited number of individual samples are available. Therefore, it is important to develop a method to identify genes associated with outcomes of interest that can control finite-sample false discovery rate (FDR) in high-dimensional data settings. ResultsThis article proposes a novel method named Grace-AKO for graph-constrained estimation (Grace), which incorporates aggregation of multiple knockoffs (AKO) with the network-constrained penalty. Grace-AKO can control FDR in finite-sample settings and improve model stability simultaneously. Simulation studies show that Grace-AKO has better performance in finite-sample FDR control than the original Grace model. We apply Grace-AKO to the prostate cancer data in The Cancer Genome Atlas (TCGA) program by incorporating prostate-specific antigen (PSA) pathways in the Kyoto Encyclopedia of Genes and Genomes (KEGG) as the prior information. Grace-AKO finally identifies 47 candidate genes associated with PSA level, and more than 75% of the detected genes can be validated. Availability and implementationWe developed an R package for Grace-AKO available at: https://github.com/mxxptian/GraceAKO Contactdoraz@hku.hk or zl2509@cumc.columbia.edu

bioinformatics↗

bioRxiv2022

Multiethnic Polygenic Risk Prediction in Diverse Populations through Transfer Learning

Polygenic risk scores (PRS) leverage the genetic contribution of an individuals genotype to a complex trait by estimating disease risk. Traditional PRS prediction methods are predominantly for European population. The accuracy of PRS prediction in non-European populations is diminished due to much smaller sample size of genome-wide association studies (GWAS). In this article, we introduced a novel method to construct PRS for non-European populations, abbreviated as TL-Multi, by conducting transfer learning framework to learn useful knowledge from European population to correct the bias for non-European populations. We considered non-European GWAS data as the target data and European GWAS data as the informative auxiliary data. TL-Multi borrows useful information from the auxiliary data to improve the learning accuracy of the target data while preserving the efficiency and accuracy. To demonstrate the practical applicability of the proposed method, we applied TL-Multi to predict the risk of systemic lupus erythematosus (SLE) in Asian population and the risk of asthma in Indian population by borrowing information from European population. TL-Multi achieved better prediction accuracy than the competing methods including Lassosum and meta-analysis in both simulations and real applications.

genetics↗