Deficiency of primate-specific SSX1 induced asthenoteratozoospermia in infertile men and cynomolgus monkey and tree shrew models.

Primate-specific genes (PSGs) tend to be expressed in the brain and testis. This phenomenon is consistent with brain evolution in primates but is seemingly contradictory to the similarity of spermatogenesis among mammals. Here, using whole-exome sequencing, we identified deleterious variants of X-linked SSX1 in six unrelated men with asthenoteratozoospermia. SSX1 is a PSG expressed predominantly in the testis, and the SSX family evolutionarily expanded independently in rodents and primates. As the mouse model could not be used for studying SSX1, we used a non-human primate model and tree shrews, which are phylogenetically similar to primates, to knock down (KD) Ssx1 expression in the testes. Consistent with the phenotype observed in humans, both Ssx1-KD models exhibited a reduced sperm motility and abnormal sperm morphology. Further, RNA sequencing indicated that Ssx1 deficiency influenced multiple biological processes during spermatogenesis. Collectively, our experimental observations in humans and cynomolgus monkey and tree shrew models highlight the crucial role of SSX1 in spermatogenesis. Notably, three of the five couples who underwent intra-cytoplasmic sperm injection treatment achieved a successful pregnancy. This study provides important guidance for genetic counseling and clinical diagnosis and, significantly, describes the approaches for elucidating the functions of testis-enriched PSGs in spermatogenesis.

Systematic evaluation of the effect of polyadenylation signal variants on the expression of disease-associated genes.

Single nucleotide variants (SNVs) within polyadenylation signals (PASs), a specific six-nucleotide sequence required for mRNA maturation, can impair RNA-level gene expression and cause human diseases. However, there is a lack of genome-wide investigation and systematic confirmation tools for identifying PAS variants. Here, we present a computational strategy to integrate the most reliable resources for discovering distinct genomic features of PAS variants and also develop a credible and convenient experimental tool to validate the effect of PAS variants on expression of disease-associated genes. This approach will greatly accelerate the deciphering of PAS variation-related human diseases.

Blood leukocytes as a non-invasive diagnostic tool for thyroid nodules: a prospective cohort study.

BACKGROUND: Thyroid nodule (TN) patients in China are subject to overdiagnosis and overtreatment. The implementation of existing technologies such as thyroid ultrasonography has indeed contributed to the improved diagnostic accuracy of TNs. However, a significant issue persists, where many patients undergo unnecessary biopsies, and patients with malignant thyroid nodules (MTNs) are advised to undergo surgery therapy. METHODS: This study included a total of 293 patients diagnosed with TNs. Differential methylation haplotype blocks (MHBs) in blood leukocytes between MTNs and benign thyroid nodules (BTNs) were detected using reduced representation bisulfite sequencing (RRBS). Subsequently, an artificial intelligence blood leukocyte DNA methylation (BLDM) model was designed to optimize the management and treatment of patients with TNs for more effective outcomes. RESULTS: The DNA methylation profiles of peripheral blood leukocytes exhibited distinctions between MTNs and BTNs. The BLDM model we developed for diagnosing TNs achieved an area under the curve (AUC) of 0.858 in the validation cohort and 0.863 in the independent test cohort. Its specificity reached 90.91% and 88.68% in the validation and independent test cohorts, respectively, outperforming the specificity of ultrasonography (43.64% in the validation cohort and 47.17% in the independent test cohort), albeit with a slightly lower sensitivity (83.33% in the validation cohort and 82.86% in the independent test cohort) compared to ultrasonography (97.62% in the validation cohort and 100.00% in the independent test cohort). The BLDM model could correctly identify 89.83% patients whose nodules were suspected malignant by ultrasonography but finally histological benign. In micronodules, the model displayed higher specificity (93.33% in the validation cohort and 92.00% in the independent test cohort) and accuracy (88.24% in the validation cohort and 87.50% in the independent test cohort) for diagnosing TNs. This performance surpassed the specificity and accuracy observed with ultrasonography. A TN diagnostic and treatment framework that prioritizes patients is provided, with fine-needle aspiration (FNA) biopsy performed only on patients with indications of MTNs in both BLDM and ultrasonography results, thus avoiding unnecessary biopsies. CONCLUSIONS: This is the first study to demonstrate the potential of non-invasive blood leukocytes in diagnosing TNs, thereby making TN diagnosis and treatment more efficient in China.

Global level of methylation in the sea lamprey (jawless vertebrate) genome is intermediate between invertebrate and jawed vertebrate genomes.

In eukaryotes, cytosine methylation is a primary heritable epigenetic modification of the genome that regulates many cellular processes. In invertebrate, methylated cytosine generally located on specific genomic elements (e.g., gene bodies and silenced repetitive elements) to show a "mosaic" pattern. While in jawed vertebrate (teleost and tetrapod), highly methylated cytosine located genome-wide but only absence at regulatory regions (e.g., promoter and enhancer). Many studies imply that the evolution of DNA methylation reprogramming may have helped the transition from invertebrates to jawed vertebrates, but the detail remains largely elusive. In this study, we used the whole-genome bisulfite-sequencing technology to investigate the genome-wide methylation in three tissues (heart, muscle, and sperm) from the sea lamprey, an extant agnathan (jawless) vertebrate. Strikingly, we found that the methylation level of the sea lamprey is very similar to that in sea urchin (a deuterostome) and sea squirt (a chordate) invertebrates. In sum, the global pattern in sea lamprey is intermediate methylation level (around 30%), that is higher than methylation level in the genomes of pre-bilaterians and protostomes (1%-10%), but lower than methylation level appeared in jawed vertebrates (around 70%, teleost and tetrapod). We anticipate that, in addition to genetic dynamics such as genome duplications, epigenetic dynamics such as global methylation reprograming was also orchestrated toward the emergence and evolution of vertebrates.

Anciently duplicated genes continuously recruited to heart expression in vertebrate evolution are associated with heart chamber increase.

Although gene/genome duplications in the early stage of vertebrates have been thought to provide major resources of raw genetic materials for evolutionary innovations, it is unclear whether they continuously contribute to the evolution of morphological complexity during the course of vertebrate evolution, such as the evolution from two heart chambers (fishes) to four heart chambers (mammals and birds). We addressed this issue by our heart RNA-Seq experiments combined with published data, using 13 vertebrates and one invertebrate (sea squirt, as an outgroup). Our evolutionary transcriptome analysis showed that number of ancient paralogous genes expressed in heart tends to increase with the increase of heart chamber number along the vertebrate phylogeny, in spite that most of them were duplicated at the time near to the origin of vertebrates or even more ancient. Moreover, those paralogs expressed in heart exert considerably different functions from heart-expressed singletons: the former are functionally enriched in cardiac muscle and muscle contraction-related categories, whereas the latter play more basic functions of energy generation like aerobic respiration. These findings together support the notion that recruiting anciently paralogous genes that are expressed in heart is associated with the increase of chamber number in vertebrate evolution.

Comprehensive analysis reveals distinct mutational signature and its mechanistic insights of alcohol consumption in human cancers.

Alcohol consumption is a critical risk factor for multiple types of cancer. A genome can be attacked and acquire numerous somatic mutations in the environment of alcohol exposure. Mutational signature has the capacity illustrating the complex somatic mutation patterns in cancer genome. Recent studies have discovered distinct mutational signatures associating with alcohol consumption in liver and esophageal cancers. However, their prevalence among diverse cancers, impact of genetic background and origin of alcohol-induced mutational signatures remain unclear. By a comprehensive bioinformatics analysis on somatic mutations from patients of four cancer types with drinking information, we identified nine mutational signatures (signatures B-J), among which signature J (similar to COSMIC signature 16) was distinctive to alcohol drinking. Signature J was associated with HNSC, ESCA and LIHC but not PAAD. Interestingly, patients with mutated allele rs1229984 in ADH1B had lower level of signature J while mutated allele rs671 in ALDH2 exhibited higher signature J abundance, suggesting acetaldehyde is one cause of signature J. Intriguingly, somatic mutations of three potential cancer driver genes (TP53, CUL3 and NSD1) were found the critical contributors for increased mutational load of signature J in alcohol consumption patients. Furthermore, signature J was enriched with early accumulated clonal mutations compared to mutations derived from late tumor growth. This study systematically characterized alcohol-related mutational signature and indicated mechanistic insights into the prevalence, origin and gene-environment interaction regarding the risk oncogenic mutations associated with alcohol intake.

Early detection and stratification of lung cancer aided by a cost-effective assay targeting circulating tumor DNA (ctDNA) methylation.

BACKGROUND: Detection of lung cancer at earlier stage can greatly improve patient survival. We aim to develop, validate, and implement a cost-effective ctDNA-methylation-based plasma test to aid lung cancer early detection. METHODS: Case-control studies were designed to select the most relevant markers to lung cancer. Patients with lung cancer or benign lung disease and healthy individuals were recruited from different clinical centers. A multi-locus qPCR assay, LunaCAM, was developed for lung cancer alertness by ctDNA methylation. Two LunaCAM models were built for screening (-S) or diagnostic aid (-D) to favor sensitivity or specificity, respectively. The performance of the models was validated for different intended uses in clinics. RESULTS: Profiling DNA methylation on 429 plasma samples including 209 lung cancer, 123 benign diseases and 97 healthy participants identified the top markers that detected lung cancer from benign diseases and healthy with an AUC of 0.85 and 0.95, respectively. The most effective methylation markers were verified individually in 40 tissues and 169 plasma samples to develop LunaCAM assay. Two models corresponding to different intended uses were trained with 513 plasma samples, and validated with an independent collection of 172 plasma samples. In validation, LunaCAM-S model achieved an AUC of 0.90 (95% CI: 0.88-0.94) between lung cancer and healthy individuals, whereas LunaCAM-D model stratified lung cancer from benign pulmonary diseases with an AUC of 0.81 (95% CI: 0.78-0.86). When implemented sequentially in the validation set, LunaCAM-S enables to identify 58 patients of lung cancer (90.6% sensitivity), followed by LunaCAM-D to remove 20 patients with no evidence of cancer (83.3% specificity). LunaCAM-D significantly outperformed the blood test of carcinoembryonic antigen (CEA), and the combined model can further improve the predictive power for lung cancer to an overall AUC of 0.86. CONCLUSIONS: We developed two different models by ctDNA methylation assay to sensitively detect early-stage lung cancer or specifically classify lung benign diseases. Implemented at different clinical settings, LunaCAM models has a potential to provide a facile and inexpensive avenue for early screening and diagnostic aids for lung cancer.

Noninvasive detection of pancreatic ductal adenocarcinoma using the methylation signature of circulating tumour DNA.

BACKGROUND: Pancreatic ductal adenocarcinoma (PDAC) has the lowest overall survival rate primarily due to the late onset of symptoms and rapid progression. Reliable and accurate tests for early detection are lacking. We aimed to develop a noninvasive test for early PDAC detection by capturing the circulating tumour DNA (ctDNA) methylation signature in blood. METHODS: Genome-wide methylation profiles were generated from PDAC and nonmalignant tissues and plasma. Methylation haplotype blocks (MHBs) were examined to discover de novo PDAC markers. They were combined with multiple cancer markers and screened for PDAC classification accuracy. The most accurate markers were used to develop PDACatch, a targeted methylation sequencing assay. PDACatch was applied to additional PDAC and healthy plasma cohorts to train, validate and independently test a PDAC-discriminating classifier. Finally, the classifier was compared and integrated with carbohydrate antigen 19-9 (CA19-9) to evaluate and maximize its accuracy and utility. RESULTS: In total, 90 tissues and 546 plasma samples were collected from 232 PDAC patients, 25 chronic pancreatitis (CP) patients and 323 healthy controls. Among 223 PDAC cases with known stage information, 43/119/38/23 cases were of Stage I/II/III/IV. A total of 171 de novo PDAC-specific markers and 595 multicancer markers were screened for PDAC classification accuracy. The top 185 markers were included in PDACatch, from which a 56-marker classifier for PDAC plasma was trained, validated and independently tested. It achieved an area under the curve (AUC) of 0.91 in both the validation (31 PDAC, 26 healthy; sensitivity = 84%, specificity = 89%) and independent tests (74 PDAC, 65 healthy; sensitivity = 82%, specificity = 88%). Importantly, the PDACatch classifier detected CA19-9-negative PDAC plasma at sensitivities of 75 and 100% during the validation and independent tests, respectively. It was more sensitive than CA19-9 in detecting Stage I (sensitivity = 80 and 68%, respectively) and early-stage (Stage I-IIa) PDAC (sensitivity = 76 and 70%, respectively). A combinatorial classifier integrating PDACatch and CA19-9 outperformed (AUC=0.94) either PDACatch (0.91) or CA19-9 (0.89) alone (p < 0.001). CONCLUSIONS: The PDACatch assay demonstrated high sensitivity for early PDAC plasma, providing potential utility for noninvasive detection of early PDAC and indicating the effectiveness of methylation haplotype analyses in discovering robust cancer markers.

Ancestral transcriptome inference based on RNA-Seq and ChIP-seq data.

With the help of high-throughput NGS (next-generation sequencing) technologies, ancestral transcriptome reconstruction is helpful to understand the complexity of transcriptional regulatory systems that underlies the evolution of multiple cellular metazoans with sophisticated functions and distinctive morphologies. To this end, we report a new method of ancestral state inference. The new method used Ornstein-Uhlenbeck (OU) model, which is more biologically realistic, to replace the Brownian motion (BM) model and is suitable for multi-transcriptome data. Implemented in the free R package, AnceTran is specially designed for RNA-seq and ChIP-seq data, which is feasible. It should be noticed that our work will be integrated to a unified, statistically-sound phylogenetic framework to study the evolution of many other molecular phenomes such as proteomics, chromatin accessibility, methylation status, and metabolomics. We exemplify our method by a case study, using the ChIP-seq binding data of three liver-specific transcription factors and the RNA-seq liver expression data in four closely related mice species, and some technical issues are discussed.

Olfactomedin domain-containing proteins: evolution, functional divergence, expression patterns and damaging SNPs.

Olfactomedin domain-containing proteins appear to facilitate neurodevelopment, cell adhesion, intercellular interactions, and protein-protein interactions, and the disruption of their expression will lead to dramatic developmental perturbations and lethality. The aim of the present work was to study how these genes evolved in metazoans and diverged after their duplication as well as to characterize their expression profiles and detrimental mutations. We conducted an exhaustive survey of olfactomedin domain-containing genes in genomic databases, identifying 235 olfactomedin-like (OLF) proteins in 29 representative species covering all the main metazoan lineages. Phylogenetic analyses allowed us to define nine different subfamilies of OLF genes, and subfamily IX, which specifically includes two immunoglobulin domains, was identified for the first time in arthropods. Functional divergence analysis suggested that the function of this arthropod-specific OLF subfamily might have diverged from that of other subfamilies. Expression pattern analysis of OLF genes in humans and rats showed that human OLF genes tended to be highly expressed in the brain, while rat OLF genes were inclined to be expressed in the ovary and brain. We used the SIFT and PolyPhen servers in dbNSFP to distinguish deleterious mutations from neutral mutations for each member of the OLF gene family. The results showed that OLFML2B contains the most destructive SNPs (up to 61), while none of the mutations in OLFM2, OLFM4 and LPHN2 were predicted to be harmful. Taken together, these findings may not only enhance understanding of the phylogenetic relationships of the OLF family but also aid future studies on OLF protein regulation of nervous system development and immune function.

Phosphotyrosine signalling and the origin of animal multicellularity.

The evolution of multicellular animals (i.e. metazoans) from a unicellular ancestor is one of the most important yet least understood evolutionary transitions. Historically, given its indispensable functions in intercellular communication and exclusive presence in metazoans, phosphotyrosine (pTyr) signalling was considered a metazoan-specific evolutionary innovation that might have contributed to the origin of metazoan multicellularity. However, recent studies have led to a new understanding of pTyr signalling evolution and its role in the metazoan origin. Sequence analyses have unravelled a much earlier emergence of pTyr signalling in eukaryotic evolution. Even so, several distinct properties of holozoan pTyr signalling may have paved the way for a hypothesized functional transition of pTyr signalling at the multicellular origin, from environmental sensing to intercellular communication, and for it to evolve as a powerful intercellular signalling system for multicellularity. Biochemical analyses of premetazoan pTyr signalling components have further revealed the premetazoan origin of many key features of metazoan pTyr signalling, and the metazoan establishment of others, including the Csk-mediated negative regulation of the activity of Src, a conserved tyrosine kinase in the Holozoa. Finally, potential future directions are discussed, with a stress on the biological functions of premetazoan pTyr signalling via newly developed gene manipulation tools in non-animal holozoans.

Brownian model of transcriptome evolution and phylogenetic network visualization between tissues.

While phylogenetic analysis of transcriptomes of the same tissue is usually congruent with the species tree, the controversy emerges when multiple tissues are included, that is, whether species from the same tissue are clustered together, or different tissues from the same species are clustered together. Recent studies have suggested that phylogenetic network approach may shed some lights on our understanding of multi-tissue transcriptome evolution; yet the underlying evolutionary mechanism remains unclear. In this paper we develop a Brownian-based model of transcriptome evolution under the phylogenetic network that can statistically distinguish between the patterns of species-clustering and tissue-clustering. Our model can be used as a null hypothesis (neutral transcriptome evolution) for testing any correlation in tissue evolution, can be applied to cancer transcriptome evolution to study whether two tumors of an individual appeared independently or via metastasis, and can be useful to detect convergent evolution at the transcriptional level.

TreeExp1.0: R Package for Analyzing Expression Evolution Based on RNA-Seq Data.

Recent innovation of RNA-seq technology has shed insightful light on the transcriptomic evolution studies, especially on researches of tissue-specific expression evolution. Phylogenetic analysis of transcriptome data may help to identify causal gene expression differences underlying the evolutionary changes in morphological, physiological, and developmental characters of interest. However, there is a deficiency of software to phylogenetically analyze transcriptome data. To address this need, we have developed an R package TreeExp that can perform comparative expression evolution analysis based on RNA-seq data, which includes optimized input formatting, normalization, pairwise expression distance estimation, expression character tree inference, and preliminary expression phylogenetic network analysis. TreeExp also enables user to map expression distance onto a customized phylogenetic tree. By applying TreeExp on two cases of mammalian gene expression evolution, we observed that (1) expression trees of brain and testis are largely consistent with known mammalian species tree with minor discrepancies; (2) intertissues expression divergences (brain and testis) are more substantial than interspecies expression divergences across mammalian species; and (3) expression pattern of gene modules related to nervous system development exhibits specific expression pattern in brain of primates compared to housekeeping genes. These tissue-specific expression patterns might give insights underlying evo-devo mechanisms of complex organisms. TreeExp is released under the GPL v3 open source license, and its current stable version 1.0 is freely available at the Github developer site (https://github.com/hr1912/TreeExp).

Asymmetric evolution of human transcription factor regulatory networks.

Changes in cis or trans regulatory regions are the major driving forces that underlie the evolution of gene expression. Transcription factors (TFs) are the main trans factors involved in transcriptional regulation. Here, we studied the divergence of upstream and downstream regulatory networks between duplicate TFs in light of the Encyclopedia of DNA Elements project. We found that the divergence of upstream regulatory networks was generally smaller than the divergence of downstream regulatory networks. Further analysis showed that the downstream regulatory circuits of duplicate TFs evolve faster in the early stage than the late stage after gene duplication. Upstream regulatory circuits are generally more conserved than downstream regulatory circuits in the early stage and in small TF families. Our results indicate the asymmetric evolution of upstream and downstream regulatory circuits between duplicate TFs, which suggest that after gene duplication, human TF families tend to evolve asymmetrically between coding regions and promoter regions.

An update of DIVERGE software for functional divergence analysis of protein family.

DIVERGE is a software system for phylogeny-based analyses of protein family evolution and functional divergence. It provides a suite of statistical tools for selection and prioritization of the amino acid sites that are responsible for the functional divergence of a gene family. The synergistic efforts of DIVERGE and other methods have convincingly demonstrated that the pattern of rate change at a particular amino acid site may contain insightful information about the underlying functional divergence following gene duplication. These predicted sites may be used as candidates for further experiments. We are now releasing an updated version of DIVERGE with the following improvements: 1) a feasible approach to examining functional divergence in nearly complete sequences by including deletions and insertions (indels); 2) the calculation of the false discovery rate of functionally diverging sites; 3) estimation of the effective number of functional divergence-related sites that is reliable and insensitive to cutoffs; 4) a statistical test for asymmetric functional divergence; and 5) a new method to infer functional divergence specific to a given duplicate cluster. In addition, we have made efforts to improve software design and produce a well-written software manual for the general user.

Age distribution patterns of human gene families: divergent for Gene Ontology categories and concordant between different subcellular localizations.

The age distribution of gene duplication events within the human genome exhibits two waves of duplications along with an ancient component. However, because of functional constraint differences, genes in different functional categories might show dissimilar retention patterns after duplication. It is known that genes in some functional categories are highly duplicated in the early stage of vertebrate evolution. However, the correlations of the age distribution pattern of gene duplication between the different functional categories are still unknown. To investigate this issue, we developed a robust pipeline to date the gene duplication events in the human genome. We successfully estimated about three-quarters of the duplication events within the human genome, along with the age distribution pattern in each Gene Ontology (GO) slim category. We found that some GO slim categories show different distribution patterns when compared to the whole genome. Further hierarchical clustering of the GO slim functional categories enabled grouping into two main clusters. We found that human genes located in the duplicated copy number variant regions, whose duplicate genes have not been fixed in the human population, were mainly enriched in the groups with a high proportion of recently duplicated genes. Moreover, we used a phylogenetic tree-based method to date the age of duplications in three signaling-related gene superfamilies: transcription factors, protein kinases and G-protein coupled receptors. These superfamilies were expressed in different subcellular localizations. They showed a similar age distribution as the signaling-related GO slim categories. We also compared the differences between the age distributions of gene duplications in multiple subcellular localizations. We found that the distribution patterns of the major subcellular localizations were similar to that of the whole genome. This study revealed the whole picture of the evolution patterns of gene functional categories in the human genome.

Circulating cell-free DNA-based multi-cancer early detection.

Patients benefit considerably from early detection of cancer. Existing single-cancer tests have various limitations, which could be effectively addressed by circulating cell-free DNA (cfDNA)-based multi-cancer early detection (MCED). With sensitive detection and accurate localization of multiple cancer types at a very low and fixed false-positive rate (FPR), MCED has great potential to revolutionize early cancer detection. Herein, we review state-of-the-art approaches for cfDNA-based MCED and their limitations and discuss both technical and clinical challenges in the development and application of MCED tests. Given the constant improvements in technology and understanding of cancer biology, we propose that a cfDNA-based targeted sequencing assay that integrates multimodal features should be optimized for MCED.

MeIS: DNA methylation-based immune response signatures for thyroid nodule diagnostics.

CONTEXT: Accurately distinguishing between benign thyroid nodules (BTNs) and papillary thyroid cancers (PTCs) with current conventional methods poses a significant challenge. OBJECTIVE: We identify DNA methylation markers of immune response-related genes for distinguishing BTNs and PTCs. METHODS: In this study, we analyzed a public reduced representative bisulfite sequencing (RRBS) dataset and revealed distinct methylation patterns associated with immune signals in PTCs and BTNs. Based on these findings, we developed a diagnostic classifier named as the Methylation-based Immune Response Signature (MeIS), which was composed of fifteen DNA methylation markers associated with immune response-related genes. We validated the MeIS's performance in two independent cohorts: ZS's retrospective cohort (50 PTC and 18 BTN surgery-leftover samples) and ZS's preoperative cohort (31 PTC and 30 BTN fine-needle aspiration (FNA) samples). RESULTS: The MeIS classifier demonstrated significant clinical promise, achieving AUCs of 0.96, 0.98, 0.89 and 0.90 in the training set, validation set, ZS's retrospective cohort, and ZS's preoperative cohort, respectively. For the cytologically indeterminate thyroid nodules, in the ZS's retrospective cohort, MeIS exhibited a sensitivity of 91% and a specificity of 82%; in the ZS's preoperative cohort, MeIS achieved a sensitivity of 84% and a specificity of 74%. Additionally, combining MeIS and BRAFV600E detection improved the detecting performance of cytologically indeterminate thyroid nodules, yielding sensitivities of 98% and 87%, and specificities of 82% and 74% in the ZS's retrospective cohort and ZS's preoperative cohort, respectively. CONCLUSIONS: The fifteen markers we identified can be employed to improve the diagnostic of cytologically indeterminate thyroid nodules.

Early detection and prognosis evaluation for hepatocellular carcinoma by circulating tumour DNA methylation: A multicentre cohort study.

BACKGROUND: Early diagnosis of hepatocellular carcinoma (HCC) can significantly improve patient survival. We aimed to develop a blood-based assay to aid in the diagnosis, detection and prognostic evaluation of HCC. METHODS: A three-phase multicentre study was conducted to screen, optimise and validate HCC-specific differentially methylated regions (DMRs) using next-generation sequencing and quantitative methylation-specific PCR (qMSP). RESULTS: Genome-wide methylation profiling was conducted to identify DMRs distinguishing HCC tumours from peritumoural tissues and healthy plasmas. The twenty most effective DMRs were verified and incorporated into a multilocus qMSP assay (HepaAiQ). The HepaAiQ model was trained to separate 293 HCC patients (Barcelona Clinic Liver Cancer (BCLC) stage 0/A, 224) from 266 controls including chronic hepatitis B (CHB) or liver cirrhosis (LC) (CHB/LC, 96), benign hepatic lesions (BHL, 23), and healthy controls (HC, 147). The model achieved an area under the curve (AUC) of 0.944 with a sensitivity of 86.0% in HCC and a specificity of 92.1% in controls. Blind validation of the HepaAiQ model in a cohort of 523 participants resulted in an AUC of 0.940 with a sensitivity of 84.4% in 205 HCC cases (BCLC stage 0/A, 167) and a specificity of 90.3% in 318 controls (CHB/LC, 100; BHL, 102; HC, 116). When evaluated in an independent test set, the HepaAiQ model exhibited a sensitivity of 70.8% in 65 HCC patients at BCLC stage 0/A and a specificity of 89.5% in 124 patients with CHB/LC. Moreover, HepaAiQ model was assessed in paired pre- and postoperative plasma samples from 103 HCC patients and correlated with 2-year patient outcomes. Patients with high postoperative HepaAiQ score showed a higher recurrence risk (Hazard ratio, 3.33, p < .001). CONCLUSIONS: HepaAiQ, a noninvasive qMSP assay, was developed to accurately measure HCC-specific DMRs and shows great potential for the diagnosis, detection and prognosis of HCC, benefiting at-risk populations.