Biology subjects

Li, D.

Publications and source records attributed to Li, D..

At least 397 records · Page 22Linked to original sources

bioRxiv2022

From multi-omics data to the cancer druggable genediscovery: a novel machine learning-based approach

The development of targeted drugs allows precision medicine in cancer treatment and achieving optimal targeted therapies. Accurate identification of cancer drug genes is helpful to strengthen the understanding of targeted cancer therapy and promote precise cancer treatment. However, rare cancer-druggable genes have been found due to the multi-omics datas diversity and complexity. This study proposes DF-CAGE, a novel machine learning-based method for cancer-druggable gene discovery. DF-CAGE integrated the somatic mutations, copy number variants, DNA methylation, and RNA-Seq data across ~10000 TCGA profiles to identify the landscape of the cancer-druggable genes. We found that DF-CAGE discovers the commonalities of currently known cancer-druggable genes from the perspective of multi-omics data and achieved excellent performance on OncoKB, Target, and Drugbank data sets. Among the ~20,000 protein-coding genes, DF-CAGE pinpointed 465 potential cancer-druggable genes. We found that the candidate cancer druggable genes (CDG-genes) are clinically meaningful and can be divided into highly reliable, reliable, and potential gene sets. Finally, we analyzed the contribution of the omics data to the identification of druggable genes. We found that DF-CAGE reports druggable genes mainly based on the CNAs data, the gene rearrangements, and the mutation rates in the population. These findings may enlighten the study and development of new drugs in the future.

bioinformatics↗

bioRxiv2022

Local fat content determines global and local stiffness in fatty livers

Fat accumulation during liver steatosis precedes inflammation and fibrosis in fatty liver diseases, and is associated with disease progression. Despite a large body of evidence that liver mechanics play a major role in liver disease progression, the effect of fat accumulation by itself on liver mechanics remains unclear. Thus, we conducted ex vivo studies of liver mechanics in rodent models of simple steatosis to isolate and examine the mechanical effects of intrahepatic fat accumulation, and found that fat accumulation softens the liver. Using a novel adaptation of microindentation to permit association of local mechanics with microarchitectural features, we found evidence that the softening of fatty liver results from local softening of fatty regions rather than uniform softening of the liver. These results suggest that fat accumulation itself exerts a softening effect on liver tissue. This, along with the localized heterogeneity of softening within the liver, has implications in what mechanical mechanisms are involved in the progression of liver steatosis to more severe pathologies and disease. Finally, the ability to examine and associate local mechanics with microarchitectural features is potentially applicable to the study of the role of heterogeneous mechanical microenvironments in both other liver pathologies and other organ systems.

physiology↗

bioRxiv2022

REST: A method for restoring signals and revealing individual macromolecule states in cryo-ET

Cryo-electron tomography (cryo-ET) is widely used to explore the 3D density of biomacromolecules. However, the heavy noise and missing wedge effect prevent directly visualizing and analyzing the 3D reconstructions. Here, we introduced REST, a deep learning strategy-based method to establish the relationship between low-quality and high-quality density and transfer this knowledge to restore signals in cryo-ET. Experimental results on purified ribosome and recombinant nucleosome datasets showed that REST had outstanding performance in denoising and compensating the missing wedge. The application in dynamic nucleosome structures suggests that REST has the capability to reveal individual macromolecules which present different conformations without subtomogram averaging. Moreover, REST could greatly improve the reliability of particle picking. These advantages enable REST to be a powerful tool for the straightforward interpretation of target macromolecules by visual inspection of the density and of a broad range of other applications in cryo-ET, such as segmentation, particle picking, and subtomogram averaging.

biophysics↗

bioRxiv2022

ribosomal DNA instability as a potential cause of karyotype evolution

Karyotype refers to the configuration of the genome into a set of chromosomes. The karyotype difference between species is expected to impede various biological processes, such as chromosome segregation or meiotic chromosome pairing, potentially contributing to incompatibility. Karyotypes can rapidly change between closely related species and even among populations of the same species. However, the forces driving karyotype evolution are poorly understood. Here we describe a unique karyotype of a D. melanogaster strain isolated from the Seychelles archipelago. This strain has lost the ribosomal DNA (rDNA) locus on the X chromosome. Because the Y chromosome is the only other rDNA-bearing chromosome, all females carry at least one Y chromosome as the source of rDNA. Interestingly, we found that the strain also carries a truncated Y chromosome (YS) that is stably maintained in the population despite its inability to support male fertility. Our modeling and cytological analysis suggest that the Y chromosome has a larger negative impact on female fitness than the YS chromosome. Moreover, we generated an independent strain that lacks X rDNA and has a karyotype of XXY females and XY males. This strain quickly evolved multiple karyotypes: two new truncated Y chromosomes (similar to YS), as well as two independent X chromosome fusions that contain the Y-derived rDNA fragment, eliminating females dependence on the Y chromosome. Considering that Robertsonian fusions frequently occur at rDNA loci in humans, we propose that rDNA loci instability is a driving force of karyotype evolution.

genetics↗

bioRxiv2022

Flavor Classification/Categorization and Differential Toxicity of Oral Nicotine Pouches (ONPs) in Lung Epithelial Cells

The prevalence of flavored tobacco product usage amongst youth in the United States is partly due to the emergence of non-combustible nicotine-containing products (NCNPs), including oral nicotine pouches (ONPs) and smokeless tobacco products. ONPs are available in various different flavors (mint, fruity, tobacco, dessert, citrus, coffee, wintergreen, and berry) and may use either Tobacco-Derived Nicotine (TDN) or Tobacco-Free Nicotine (TFN). Currently, several brands of ONPs are sold in the U.S and comprise a significant portion of NCNP sales in the U.S. There is a growing concern that flavored ONPs may not only induce oral health effects, but may also induce systemic toxic effects due to nicotine and other ONP byproducts being absorbed into systemic circulation through the oral mucosa. These byproducts can act locally on other tissues and may potentially cause redox dysregulation and heightened inflammatory responses systemically in the respiratory, cardiovascular, and/or renal systems. Hence, we determined the effects of flavored ONPs from four of the most widely sold brands in the U.S in inducing toxicological effects on the respiratory epithelium. Prior to analyzing the effects ONPs, we first classified ONPs sold in the US based on their flavor and the flavor category to which they belong to using a wheel diagram. Subsequently, using human bronchial epithelial cells (16-HBE and BEAS-2B) exposed to extracts of flavored ONPs, we assessed the levels of ONP-induced inflammatory cytokine release (IL-6 and IL-8), cellular Reactive Oxygen Species (ROS) production, and cytotoxicity in the airway epithelium. Our data showed that cells exposed to the lowest concentration treatments showed increased cytotoxicity, differential cellular ROS production, and proinflammatory cytokine release. The most striking response was observed among cells treated with the spearmint ONP, whereas ONPs containing original tobacco and fruity flavors showed varied levels of ROS release in 16-HBE cells. Our data suggest that flavored ONPs are unsafe and likely to cause systemic and local toxicological responses during chronic usage. Our study is a part of ongoing efforts to use in vitro, ex-vivo, and in vivo systems to understand how the usage of various flavored ONPs may cause both oral and pulmonary toxicity, and impact human periodontal health.

pharmacology and toxicology↗

bioRxiv2022

Human skin specific long noncoding RNA HOXC13-AS regulates epidermal differentiation by interfering with Golgi-ER retrograde transport

After a skin injury, keratinocytes switch from a state of homeostasis to one of regeneration leading to the reconstruction of the epidermal barrier. The regulatory mechanism of gene expression underpinning this key switch during human skin wound healing is enigmatic. Long noncoding RNAs (lncRNAs) constitute a new horizon in the understanding of the regulatory programmes encoded in the mammalian genome. By comparing the transcriptome of an acute human wound and skin from the same donor as well as keratinocytes isolated from these paired tissue samples, we generated a list of lncRNAs showing changed expression in keratinocytes during wound repair. Our study focused on HOXC13-AS, a recently evolved human lncRNA specifically expressed in epidermal keratinocytes, and we found that its expression was temporally downregulated during wound healing. In line with its enrichment in suprabasal keratinocytes, HOXC13-AS was found to be increasingly expressed during keratinocyte differentiation, but its expression was reduced by EGFR signalling. After HOXC13-AS knockdown or overexpression in human primary keratinocytes undergoing differentiation induced by cell suspension or calcium treatment and in organotypic epidermis, we found that keratinocyte differentiation was promoted, while the cell inflammatory response was suppressed. Moreover, RNA pull-down assays followed by mass spectrometry and RNA immunoprecipitation analysis revealed that mechanistically HOXC13-AS sequestered the coat complex subunit alpha (COPA) protein and interfered with Golgi-to-endoplasmic reticulum (ER) molecular transport, resulting in ER stress and enhanced keratinocyte differentiation. In summary, we identified HOXC13-AS as a crucial regulator of human epidermal differentiation.

molecular biology↗

bioRxiv2022

Evolutionary survival strategies of the female giant panda: optimizing energy resources and expenditure prior to pregnancy by postponing corpus luteum reactivation

The giant panda reproductive physiology shows important similarities with at least six of the eight existing bear species: the occurrence of diapause and/or pseudopregnancy is commonly described in bears. Nevertheless, significant differences including the earlier breeding season with - in general - a single estrus, a shorter delay of implantation and a more variable birth season, indicate an evolutionary adaptation from the seasonal reproductive traits described in hibernating bear species. In this study we aimed to determine true pregnancy length for giant pandas and to open the discussion on the peculiarities of giant panda reproductive biology, more specifically focusing on the rationale behind their short-seasoned reproductive cycle compared to the other bear species. For this purpose, metabolic (body weight and fecal output) profiles were matched with endocrine changes, mainly urinary progesterone metabolites, in 5 pregnant, 8 non-birth and 6 pseudopregnant cycles of 6 female giant pandas. Pregnancy in giant pandas lasts only 42 days from early reactivation of the corpora lutea (CLs) until birth. In addition, our findings urged the need to redefine the generally accepted biphasic progesterone profile into a triphasic primary progesterone rise (corpus luteum dormancy (CLD) I, II and III) prior to entering the active luteal phase (= secondary progesterone rise). Two episodes of progesterone increase (CLDII: 81.20 {+/-} 5.85 days versus CLDIII: 60.80 {+/-} 3.83 days prior to birth for pregnant cycles) were identified, respectively corresponding to CL reactivation (74-88 days prior to birth) and implantation ({+/-} 60 days prior to birth) in other bear species. The progesterone concentration during CLDIII was higher in pregnant cycles, indicating a potential communication between maternal tissues and blastocyst(s) enhancing progesterone concentrations and thus allowing optimal priming of uterine tissues to better prepare for blastocyst reactivation/development. Compared with other bear species, giant pandas seem to shorten the active luteal phase, and thus pregnancy, by approximately 30 days by actively postponing CL reactivation. Potential mechanisms in play overruling and suppressing the evolutionary conserved photoperiodical triggers of CL reactivation are discussed while a parallel study will further elaborate on the CL dynamics during giant panda gestation.

evolutionary biology↗

bioRxiv2022

PTEN and BRCA1 tumor suppressor loss associated tumor immune microenvironment exhibits differential response to therapeutic STING pathway activation in a murine model of ovarian cancer

BackgroundHigh grade serous ovarian carcinoma (HGSC) is the most lethal gynecologic malignancy characterized by chemoresistance and high rates of recurrence. HGSC tumors display a high prevalence of tumor suppressor gene loss. Loss of BRCA1 and PTEN function due to mutations or epigenetic influence have been widely associated with variable clinical outcomes, where tumors with BRCA1 mutations exhibit increased chemosensitivity and those with PTEN mutations have been reported to exhibit chemoresistance. Given the established type 1 interferon regulatory function of BRCA1 and PTEN genes and associated contrasting T cell infiltrated and non-infiltrated tumor immune microenvironment (TIME) states, in this study we investigated the potential of Stimulator of Interferon Genes (STING) pathway activation in improving overall survival via enhancing chemotherapy response, specifically in tumors with PTEN deficiency. MethodsExpression of PTEN protein was evaluated in tissue microarrays generated using pre-treatment tumors collected from a cohort of 110 patients with HGSC. Multiplex immunofluorescence staining was performed to determine spatial profiles and density of selected lymphoid and myeloid cells. In vivo studies using the syngeneic murine HGSC cell lines, ID8-Trp53-/-; Pten-/-and ID8-Trp53-/-; Brca1-/-, were conducted to characterize the TIME and response to carboplatin chemotherapy in combination with exogenous STING activation therapy. ResultsTumors with absence of PTEN protein exhibited a significantly decreased disease specific survival and intra-epithelial CD68+ macrophage infiltration as compared to intact PTEN expression. In vivo studies demonstrated that Pten deficient ovarian cancer cells establish an immunosuppressed TIME characterized by increased proportions of M2-like macrophages, GR1+ MDSCs in the ascites, and reduced effector CD8+ cytotoxic T cell function compared to Brca1 deficient cells; further, tumors from mice injected with Pten deficient ID8 cells exhibited an aggressive behavior due to suppressive macrophage dominance in the malignant ascites. In combination with chemotherapy, exogenous STING activation resulted in longer overall survival in mice injected with Pten deficient ID8 cells, reprogrammed intraperitoneal M2-like macrophages derived from Pten deficient ascites to a M1-like phenotype and rescued CD8+ cytotoxic T cell activation. ConclusionsThis study reveals the importance of considering the influence of cancer cell intrinsic genetic alterations on the TIME for therapeutic selection. We establish the rationale for the optimal incorporation of interferon activating therapies as a novel combination strategy in PTEN deficient HGSC.

immunology↗

bioRxiv2022

Transcriptomic and functional mapping of autism associated environmental factors in developing human neurons

Research continues to identify genetic variation, environmental exposures, and their mixtures underlying different diseases and conditions. There is a need for screening methods to understand the molecular outcomes of such factors. Here, we investigate a highly efficient and multiplexable, fractional factorial experimental design (FFED) to study six environmental factors and four human induced pluripotent stem cell line derived differentiating human neural progenitors. We showcase the FFED coupled with RNA-sequencing to identify the effects of low-grade exposures to these environmental factors and analyse the results in the context of autism spectrum disorder (ASD). We performed this after five-day exposures on differentiating human neural progenitors accompanied by a layered analytical approach and detected several convergent and divergent, gene and pathway level responses. We revealed significant upregulation of pathways related to synaptic function and lipid metabolism following lead and fluoxetine exposure, respectively. The lipid changes were validated using mass spectrometry- based metabolomics after fluoxetine exposure. Our study demonstrates that the FFED can be used for multiplexed transcriptomic analyses to detect relevant pathway-level changes in human neural development caused by low-grade environmental risk factors. Future studies will require multiple cell lines with different genetic backgrounds for characterising the effects of environmental exposures in ASD.

neuroscience↗

bioRxiv2022

The Optimal Combination of Scanning Speed and Fiber-to-Stone Distance for Effective and Efficient Dusting during Holmium: YAG Laser Lithotripsy

ObjectivesTo investigate mechanistically the effects of fiber scanning speed (vfiber) and fiber tip-to-stone standoff distance (SD) on dusting efficiency during Holmium (Ho): YAG laser lithotripsy (LL) Materials and MethodsPre-soaked BegoStone samples (23 x 23 x 4 mm, W x L x H) were treated in water using a clinical Ho:YAG laser in dusting mode (0.2 J pulse energy delivered at 20 Hz frequency) at three different SDs (0.10, 0.25 and 0.50 mm) with vfiber in the range of 0 to 10 mm/s. Stone damage was quantified by optical coherence tomography, bubble dynamics were captured by high-speed imaging, and associated pressure transients were measured using a needle hydrophone. To compare photothermal ablation vs. cavitation in stone dusting, the experiments were further repeated in air (photothermal ablation only), and in water with the fiber tip advanced at a short (0.25 mm) offset distance (OSD) from a ureteroscope to mitigate the bubble collapse toward the stone surface, thus eliminating cavitation-induced damage. ResultsCompared to the craters produced by a stationary fiber, the damage troughs produced by a scanning fiber after 100 pulses were significantly larger in volume. The optimal vfiber for maximum dusting efficiency was found to be 3.5 mm/s for SD = 0.10 [~] 0.25 mm, resulting in long (17.5 mm), shallow (0.14 - 0.15 mm) and narrow (0.3 - 0.4 mm) troughs. In contrast, the maximum stone damage was produced at an optimal vfiber of 0.5 mm/s for SD = 0.50 mm, which generates much shorter (2.5 mm), yet deeper (0.35 mm) and wider (1.4 mm) troughs. Greater stone damage was produced in water than in air, especially at vfiber = 0 - 3.5 mm/s. With the scope end placed near the fiber tip, stone damage could be significantly reduced in water by 29% - 58% for SD = 0.10 mm, by 51% - 82% for SD = 0.25 mm, and by 66% - 100 % for SD = 0.50 mm, compared to those produced without the scope. Together, these findings suggest that cavitation plays an indispensable role in stone dusting by scanning treatment. Moreover, under clinically relevant vfiber (1 [~] 3 mm/s), dusting at SD = 0.5 mm (i.e., non-contact mode) may leverage higher frequency of the laser (e.g., 40 to 120 Hz) to harvest the full potential of cavitation damage while significantly reducing the procedure time, compared to its counterpart at SD = 0.1 mm (i.e., contact mode) that promotes photothermal ablation. ConclusionDusting efficiency during Ho:YAG LL may be substantially improved by utilizing the optimal combination of vfiber and SD for a given frequency.

bioengineering↗

bioRxiv2022

Clean-PIE: a novel strategy for efficiently constructing precise circRNA with thoroughly minimized immunogenicity to direct potent and durable protein expression

Translatable circular RNAs (circRNAs) are emerging as a crucial molecular format for transient protein expression, with high potential to be an alternative for linear mRNA to reshape the landscape of mRNA pharmaceutical industry. Canonical Anabaena permuted intron-exon (Ana PIE) format that developed by ORNA is an efficient method for RNA circularization, and the engineered circRNAs direct supreme protein expression in eukaryotic cells. However, recent studies revealed that this method may unavoidably result in a remain of immunogenicity in the circRNA products, albeit after thorough RNA purification. In the current study, we develop a novel strategy for efficient generation of circRNA, via the permuted T4Td introns mediated autocatalytically ribozymatic reaction mediated ligation of the flanking segment sequences that concealing in ORF or translation initiation sequence (normally equal to IRES). This strategy universally realizes around 90% circularization effectivity, and the circRNA products can be purified to around 90% purity by our new purification method, and presented thoroughly minimized immunogenicity, thus is termed "Clean-PIE". The purified circRNAs are found to direct potent and durable expression of various proteins in vitro and in vivo. The partly purified Fluc circRNA by HPLC-SEC was found to direct Fluc expression in muscle for no less than 20 days. The highly purified circRNA exhibits much stronger protein expression in vitro and in vivo, and presumed a longer duration. Additionally, the scale-up of RNA circularization with the RNA precursors from 1 L transcription revealed high circularization effectivity (around 90%) and a high productivity of the final circRNA products. Collectively, Clean-PIE is a novel circRNA platform that possesses high circularization effectivity, enabled high RNA purity and thoroughly minimized immunogenicity, as well as scaling-up accessibility and directing extreme durability of protein expression, thus has the potential to develop advanced RNA vaccines and therapeutics in pharmaceutical industrial scale.

molecular biology↗

bioRxiv2022

Life cycle exposure to cyhalofop-butyl induced reproductive toxicity toward zebrafish

Cyhalofop-butyl (CyB) is an herbicide widely used in paddy fields, which may transfer to aquatic ecosystems and cause harm to aquatic organisms. In this study, zebrafish (Danio rerio) were exposed to CyB (0.1, 1 and 10 ug/L) related to the environment throughout its adult life cycle from embryo to sexual maturity. The effects of CyB on zebrafish growth, reproduction and offspring development were studied. It was found that female spawning was inhibited and adult male fertility decreased. In addition, we detected the expression of sex steroid hormones and genes related to hypothalamus-pituitary-gonad-liver (HPGL) axis. After 150days of exposure, the hormone balance of parent zebrafish (F0) was disturbed and the concentrations of 17{beta}-estradiol (E2) and vitellogenin (VTG) in zebrafish were decreased. F1 embryos showed abnormal developmental results, including decreased heart rate, decreased body length, spontaneous motor inhibition, while the developmental abnormalities of F1 embryos were relieved when exposed to CyB-free clear water. The change of sex hormone is regulated by gene expression related to HPGL axis. These results confirmed that long-term exposure to CyB in the environmental concentrations can damage the reproductive capacity of F0 generation zebrafish by disrupting the transcription of genes related to HPGL axis, which may lead to abnormal development of F1. Overall, these data may provide a new understanding of the reproductive toxicity of zebrafish parents and offspring after long-term exposure to CyB. HighlightsO_LIEnvironmental level of CyB exposure caused gonadal impairment. C_LIO_LICyB exposure suppressed spawning ability of zebrafish. C_LIO_LICyB exposure changed the plasma hormone level of zebrafish and altered HPGL axis in both genders. C_LIO_LIParental CyB exposure led to abnormal development of F1. C_LI

pharmacology and toxicology↗

bioRxiv2022

Epithelial protein lost in neoplasm (EPLIN)-beta is a novel substrate of ornithine decarboxylase antizyme 1, mediating cellular migration

Polyamines promote cellular proliferation, and their levels are controlled by ornithine decarboxylase antizyme 1 (Az1), through the proteasome-mediated, ubiquitin-independent degradation of ornithine decarboxylase (ODC), the rate-limiting enzyme of polyamine biosynthesis. Az1-mediated degradation of other substrates such as cyclin D1, DNp73 and Mps1 also regulate cellular migration and centrosome amplification, and collectively, the currently known six Az1-substrates are all linked with tumorigenesis. To understand if Az1-mediated protein degradation might play a key role in regulating cellular process associated with tumorigenesis, we employed quantitative proteomics to identify novel Az1 substrates. In this report, we describe the identification of LIM domain and Actin-binding protein 1 (LIMA1), also known as epithelial protein lost in neoplasm (EPLIN), as a new target of Az1. Interestingly, between the two isoforms of EPLIN ( and {beta}), only EPLIN-{beta} is the substrate of Az1, degraded in a proteasome-dependent and ubiquitination-independent manner. Absence of Az1 led to elevated EPLIN-{beta} levels, which is causal to enhanced cellular migration of Az1-/- cells. Consistently, higher levels of LIMA1 expression correlated with poorer overall survival of colorectal cancer patients. Taken together, this study identifies EPLIN-{beta} as a novel Az1 substrate that regulates cellular migration.

cancer biology↗

bioRxiv2022

Connexin43 Overexpression Promotes Bone Regeneration by Osteogenesis and Angiogenesis in Rat Glucocorticoid-Induced Osteonecrosis of the Femoral Head

Glucocorticoids induced osteonecrosis of the femoral head (GIONFH) is a devastating orthopedic disease. Previous studies suggested that connexin43 (Cx43) is involved in the process of osteogenesis and angiogenesis. However, the role of Cx43 potentiates in the osteogenesis and angiogenesis of bone marrow-derived stromal stem cells (BMSCs) in GIONFH is still not investigated. In this study, BMSCs were isolated and transfected with green fluorescent protein (GFP) or the fusion gene encoding GFP and Cx43. The osteogenic differentiation of BMSCs were detected after transfected with Cx43. In addition, the migration abilities and angiogenesis of human umbilical vein endothelial cells (HUVECs) were been detected after induced by transfected BMSCs supernatants in vitro. Our results showed that Cx43 overexpression in BMSCs promoted osteogenic differentiation and angiogenesis in vitro. Finally, we established GC-ONFH rat model, then, a certain amount of transfected or controlled BMSCs were injected into the tibia of the rats. Immunohistological staining and micro-CT scanning results showed that the transplanted experiment group had significantly promoted more bone regeneration, vessel volume and the expressions of Runx2, ALP, COL I, VEGF and CD31 when compared with the effects of the negative or control groups. This study demonstrated for the first time that the Cx43 overexpression in BMSCs could promote bone regeneration as seen in the osteogenesis and angiogenesis process, suggesting that Cx43 may serve as a therapeutic gene target for GIONFH treatment.

cell biology↗

bioRxiv2022

Trans-Driver: a deep learning approach for cancer driver gene discovery with multi-omics data

Driver genes play a crucial role in the growth of cancer cells. Accurate identification of cancer driver genes is helping to strengthen the understanding of cancer pathogenesis and is conducive to the development of cancer treatment and drug-targe driver genes. However, due to the diversity and complexity of the multi-omics data, it is still challenging to identify cancer drivers.In this study, we propose Trans-Driver, a deep supervised learning method with a novel transformer network, which integrates multi-omics data to learn the differences and associations between different omics data for cancer drivers discovery. Compared with other state-of-the-art driver gene identification methods, Trans-Driver has achieved excellent performance on TCGA and CGC data Machine learning for multi-omics data integration in cancer. Among 20,000 protein-coding genes, Trans-Driver reported 185 candidate driver genes, of which 103 genes (about 55%) were included in the gold standard CGC data set. Finally, we analyzed the contribution of each feature to the identification of driver genes. We found that the integration of multi-omics data can improve the performance of our method compared with using only somatic mutation data. Through detailed analysis, we found that the candidate drivers are clinically meaningful, proving the practicability of Trans-Driver. Author summaryMany methods have been developed to identify cancer driver genes. However, most of these methods use single-omics data for cancer driver gene identification. Multi-omics-based methods for cancer driver gene identification are rare. Trans-Driver uses deep learning to process multi-omics data and learn the relationships between multi-omics data for cancer driver gene prediction. We have predicted 185 candidate cancer driver genes out of among 20,000 protein-coding genes. Also, we performed cancer driver gene prediction on 33 cancer types, and we identified the cancer driver genes corresponding to each cancer type. And, we observed that the predicted cancer driver genes were shown to have a role in cancer progression in recent studies. Our proposed method for cancer driver gene identification using multi-omics data has improved performance compared to using mutation data alone.

bioinformatics↗

bioRxiv2022

Towards generalizable prediction of antibody thermostability using machine learning on sequence and structure features

Over the last three decades, the appeal for monoclonal antibodies (mAbs) as therapeutics has been steadily increasing as evident with FDAs recent landmark approval of the 100th mAb. Unlike mAbs that bind to single targets, multispecific biologics (bsAbs) with their single-chain variable fragment (scFv) modules have garnered particular interest owing to the advantage of engaging distinct targets. Despite their exquisite specificity and affinity, the relatively poor thermostability of these scFv modules often hampers their development as a potential therapeutic drug. In recent years, engineering antibody sequences to enhance their stability by mutations has gained considerable momentum. As experimental methods for antibody engineering are time-intensive, laborious, and expensive, computational methods serve as a fast and inexpensive alternative to conventional routes. In this work, we show two machine learning methods - one with pre-trained language models (PTLM) capturing functional effects of sequence variation, and second, a supervised convolutional neural network (CNN) trained with Rosetta energetic features - to better classify thermostable scFv variants from sequence. Both these models are trained over temperature-specific data (TS50 measurements) derived from multiple libraries of scFv sequences. In this work, we show that a sufficiently simple CNN model trained with energetic features generalizes better than a pre-trained language model on out-of-distribution (blind) sequences (average Spearman correlation coefficient of 0.4 as opposed to 0.15). Further, we demonstrate that for an independent mAb with available thermal melting temperatures for 20 experimentally characterized thermostable mutations, these models trained on TS50 data could identify 18 residue positions and 5 identical amino-acid mutations showing remarkable generalizability. Our results suggest that such models can be broadly applicable for improving the biological characteristics of antibodies. Further, transferring such models for alternative physico-chemical properties of scFvs can have potential applications in optimizing large-scale production and delivery of mAbs or bsAbs.

bioinformatics↗

bioRxiv2022

PRC2 Inactivating Mutations in Cancer are Synthetic Lethal with DNMT1 Targeted Therapy via Enhanced Viral Mimicry

Polycomb Repressive Complex 2 (PRC2) establishes and maintains di- and tri-methylation at histone 3 at lysine 27 (H3K27me2/3) in the genome and plays oncogenic and tumor suppressor roles in context-dependent cancer pathogenesis. While there is clinical success of therapeutically targeting PRC2 core component, EZH2, in PRC2-dependent cancers (e.g., follicular lymphoma, epithelioid sarcoma), it remains an unmet therapeutic bottleneck in PRC2-inactivated cancer. Biallelic inactivating mutations in PRC2 core components are a hallmark feature of high-grade malignant peripheral nerve sheath tumor (MPNST), an aggressive subtype of sarcoma with poor prognosis and no effective targeted therapeutics. Using a custom RNAi-based drop out screen, we observed that PRC2-inactivation is synthetic lethal with DNA methyltransferase 1 (DNMT1) downregulation; we further observed that small molecule DNMT inhibitors (DNMTis) resulted in enhanced cytotoxicity and antitumor response in PRC2-loss cancer context in vitro and in vivo. Mechanistically, DNMTi-mediated de-repression of retrotransposons (e.g., endogenous retroviral elements (ERVs)/LTR, LINE, SINE) and gene targets is partly restricted by PRC2, which potentially contributes to limited therapeutic activity in PRC2-wild-type (wt) cancer context. In contrast, DNMTi treatment synergizes with PRC2 inactivation and cooperatively amplifies the expression of retrotransposons (e.g., ERV/LTR, LINE, SINE), and subsequent viral mimicry response that promotes robust cell death in part through PKR-dependent double stranded-RNA (dsRNA) sensing. Collectively, our observations posit DNA methylation as a safeguard against anti-tumorigenic cell fate decisions in the context of PRC2-inactivation to promote cancer pathogenesis. Further, they identified a novel targeted therapeutic strategy in PRC2-inactivated MPNST and delineated the PRC2-inactivated cancer context for future preclinical exploration and clinical investigation of DNMT1-targeted therapies in cancer. SIGNIFICANCEPRC2-inactivation drives oncogenesis in various cancers but therapeutically targeting PRC2-loss has remained challenging. Here we show that PRC2 inactivating mutations sets up a tumor context-specific liability for synthetic lethal interaction with genetic and therapeutic inhibition of DNMT1. DNMT1 inhibitor-induced cytotoxicity in PRC2-loss cancer context is accompanied by innate immune signaling signature through PKR-mediated sensing of endogenous retrotransposons. These observations posit a therapeutic window via direct anti-tumor effect by DNMT1 inhibitors in PRC2-loss cancers, and point to potentials to be combined with innovative immunotherapeutic strategies to capitalize on innate immune signaling activation.

cancer biology↗

bioRxiv2022

The dynseq genome browser track enables visualization of context-specific, dynamic DNA sequence features at single nucleotide resolution

We introduce the dynseq genome browser track, which displays DNA nucleotide characters scaled by user-specified, base-resolution scores provided in the BigWig file format. The dynseq track enables visualization of context-specific, informative genomic sequence features. We demonstrate its utility in three popular genome browsers for interpreting cis-regulatory sequence syntax and regulatory variant interpretation by visualizing nucleotide importance scores derived from machine learning models of regulatory DNA trained on protein-DNA binding and chromatin accessibility experiments.

genomics↗