Single-cell copy number alteration signature analysis reveals masked patterns and potential biomarkers for cancer

Single-cell copy number alteration signature analysis reveals masked patterns and potential biomarkers for cancer | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Single-cell copy number alteration signature analysis reveals masked patterns and potential biomarkers for cancer Xue-Song Liu, Chenxu Wu, Ziyu Tao, Wen Chen, Nan Wang, Jinyu Wang, and 11 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6304422/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 17 Nov, 2025 Read the published version in Communications Biology → Version 1 posted You are reading this latest preprint version Abstract Copy number alteration (CNA) is a major type of cancer genome alteration that drives cancer progression. CNA signature analysis can reveal underlying etiology and provide biomarkers for cancer treatment, and existing CNA signature analyzes are all performed using bulk tissue samples. However CNA usually affect large proportion of genome, and the CNA profile of bulk sample does not reflect the actual CNA profiles of the individual cancer cells of the sample, especially in tumors with high heterogeneity, such as hepatocellular carcinoma (HCC). Furthermore, the evolutionary trajectory of CNA mutational processes still remain elusive. Here we build a method to comprehensively analyze the CNA signatures of HCC from single-cell and bulk sample perspective, revealing patterns and potential noise signals from the usually performed bulk tissue CNA signature analysis. Single-cell signature analysis delineated the evolutionary trajectory of HCC CNA signatures, and different CNA signatures consistently emerge in different HCC evolution stages. Single-cell CNA signatures show robust performance in patient prognosis and drug sensitivity prediction. This work not only reveals specific considerations in analyzing CNA signature derived from bulk tissue but also depicts CNA evolution process and provides potential biomarkers for the prognosis and treatment of HCC patients. Biological sciences/Cancer/Cancer genomics Biological sciences/Computational biology and bioinformatics/Genome informatics scDNA-seq copy number signature HCC evolutionary Drug sensitivity Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Highlights Single-cell analysis reveals CNA signatures masked in bulk tissue. Single-cell analysis delineates the evolutionary trajectory of CNA signature. Small CNAs occur early and large CNAs happens late in HCC evolution. Single-cell CNA signatures show robust performance in guiding cancer clinical treatment. Introduction Genomic instability is one of the hallmarks of cancer 1 , encompassing changes ranging from single nucleotide alterations to whole chromosome modifications. Numerous studies have demonstrated the roles of single base substitutions (SBS), small insertions and deletions (INDEL), structural variations (including translocations/inversions), and copy number alterations in the process of cancer development 2 – 4 . Genomic DNA alteration signatures are recurring genomic patterns that are the imprints of mutagenic processes accumulated over the lifetime of cancer cell 5 . SBS signature analysis has been extensively studied and represents a prototype for other types of signature study 5 . While analyzing copy number signature is more intricate than SBS signatures. Pan-cancer CNA signature studies have been carried out 6 – 9 , and these studies indicate that genomic alteration signatures can provide information on mutational processes, and also serve as biomarkers for precision medicine in cancer 10 – 15 . Current research on CNA signature is predominantly focused on bulk tissue samples. In the detection of SBS, bulk tissue is capable of preserving the SBS mutation status of the majority of subclones. In contrast, CNA usually affect large proportion of genome DNA, and the final CNA profile of bulk tissue is usually not the linear combination of each subclone's CNA state (Fig. 1 as example). For tumors with low heterogeneity, sequencing of bulk tissue effectively captures the variations in copy numbers. However, for tumors with high heterogeneity, like hepatocellular carcinoma (HCC) 16 – 18 , sequencing of bulk tissue only reflects the averaged CNA across all tumor subclones 19 , 20 , failing to accurately reveal the true patterns of CNA. The models of tumor evolution have been a contentious subject in tumor biology 21 , 22 . Gao et al's model of punctuated copy number evolution suggests that copy number aberrations are primarily acquired during brief peaks of genomic crisis, followed by stable clonal expansion 23 . Guo et al 24 introduced a model of dual-phase copy number evolution in HCC, indicating the coexistence of gradual and punctuated evolution. Although these tumor evolution models have revealed the occurrence rate of CNAs during tumor progression, they have not clarified which CNA patterns or signatures appear in the early stages of a tumor and which appear in the later stages. The evolutionary trajectory of CNA signatures still remain elusive. In this study, we collected scDNA-seq data from 1222 cells of 10 HCC patients and developed a versatile tool for analyzing CNA signatures that apply to both single-cell and bulk tissue samples. This tool enabled us to systematically analyze the CNA signatures in HCC at the single-cell level for the first time. Comparing CNA signatures derived from single-cell and those from bulk samples revealed that some CNA signatures observed in bulk tissue samples might not actually exist. Additionally, we have for the first time proposed the evolutionary trajectory of CNA signatures in HCC. These studies provide new insights into the mutation processes of HCC and important information for the in-depth study of CNA signatures and their potential clinical applications. Results Single-cell analysis revealed masked CNA profiles compared with bulk tissue CNA analysis Each point mutation occurs on a single chromosome in a single cell, which gives rise to a lineage of cells bearing the same mutation. If that chromosomal locus is subsequently duplicated, any point mutation on this allele preceding the gain will subsequently be present on the two resulting allelic copies 25 , 26 . This process facilitates the accumulation of SBS without interference between mutations. Bulk tissue sequencing enables point mutation detection within a sample, reflecting the cumulative outcome of all subclones (Fig. 1 A). However CNA usually affect large proportion of genome DNA, when a CNA occurs within a single cell, the subsequent subclonal CNA further modify the original CNA profile, distorting its characteristic signature. Consequently, the CNA observed in bulk tissue is not an accurate reflection of the true CNA occurs in each subclones (Fig. 1 B). This phenomenon is particularly pronounced in tumors with high heterogeneity, such as HCC. To accurately explore the true state of CNA, we aggregated data from various HCC databases, covering different sequencing platforms and depths. Our collection comprised 1222 single-cell shallow whole-genome sequencing (sWGS) samples 24 , 27 , 160 bulk sWGS samples 28 , 178 bulk whole-genome sequencing (WGS) samples 29 , 30 , and 371 bulk SNP array samples 31 . All raw data were uniformly processed to derive the absolute copy number information of the samples (details provided in the Materials and Methods). Development of CNA signature analysis tool for both single-cell and bulk tissue samples To identify single-cell CNA signatures, we developed a novel method for the extraction of CNA signatures. This method encompasses four principal aspects of CNA: absolute copy number, segment length, segment change, and segment shape (Fig. 2 ). From these aspects, we further delineated 90 distinct features (Figure S1 ). These features were selected as hallmarks of previously reported genomic aberrations, including chromothripsis 32 , large-scale state transitions (LST) 33 , extrachromosomal circular DNA (ecDNA) 34 , and tandem duplications 35 . Following the computation of features for all samples, this feature matrix was processed using non-negative matrix factorization to identify CNA signatures (Fig. 2 ). This method has previously been employed for extracting signatures of SBS 36 , 37 . The number of signatures extracted was determined using two parameters. First is the reconstruction error and the average Frobenius reconstruction error is reported. Second is the stability of signature extraction and the cosine similarity between the extracted signatures, and average silhouette width. In the single-cell samples, a total of seven CNA signatures were identified. In the analysis of bulk samples, seven CNA signatures were identified with sWGS and WGS data. Eight CNA signatures were determined using SNP array data (Figure S2). Mutational processes for single-cell CNA signatures Following determining CNA signatures, we initially assessed the similarity among different CNA signatures. The observed similarities between CNA signatures were low (Figure S3), suggesting that each CNA signature is relatively distinct (Fig. 3 A and Figure S4). Further, we compared the CNA signatures extracted from bulk tissue sWGS, WGS, and SNP array data, with those obtained from single-cell sWGS data. Our results indicate that the CNA signatures identified in single-cell shallow sequencing data can also be detected in other data types (Figure S5A-C), thus demonstrating the stability and reliability of our proposed CNA signature extraction method. However, specific CNA signatures, such as WGS_Sig5, WGS_Sig6, SNP_Sig3, and SNP_Sig4, did not have similar counterparts in single-cell CNA signatures. This phenomenon may be attributed to several factors. On the one hand, the limited number of patients represented in the single-cell samples may have resulted in the exclusion of certain sample-specific CNA signatures from the analysis. On the other hand, these CNA signatures may represent signals specific to bulk tissues rather than single cells. Association analysis between CNA signatures and phenotypic data can reveal the underlying clinical and biological information for CNA signatures 8 , 11 , 15 . In the scSig1, we identified numerous genomic fragments smaller than 5Mb with high copy numbers, consistent with genomic variations induced by ecDNA 38 – 41 . Further associative analyses suggest these variations are attributable to ecDNA (Figure S6). Based on the analysis of representative samples of CNA signatures, we observed that the scSig3 pattern reflects characteristics of a normal genome, while the genome with the scSig4 pattern exhibits a clear triploid state, aligning with the CNS3 and CNS12 proposed by Tao et al 8 . Utilizing these analyses, we have summarized the potential etiologies of each CNA signature in single-cell data from HCC (Fig. 3 B). Single-cell CNA signature analysis reveals masked patterns in bulk tissue analysis To further investigate the differences between single-cell CNA signatures and bulk tissue CNA signatures, following the methodology of Drews et al 7 , this study simulated CNA events, including early and late whole-genome duplication (early WGD, late WGD) 42 , 43 and LST (associated with homologous recombination deficiency, HRD) 33 . The results indicate that single-cell CNA signatures can distinguish early WGD from late WGD, however bulk tissue CNA signatures fail to effectively differentiate early from late WGD events (Fig. 4 A). This finding reflects the advantag e of single-cell CNA signatures in finely distinguishing patterns of CNA. Following the genomic variation simulation method of Bartenhagen et al 44 , we simulated the formation of a series of CNA for comparative analysis, and the results also demonstrated that single-cell CNA signatures are more robust than bulk tissue CNA signatures in filtering out such simulated noise (Fig. 4 B and Figure S7). We employed the method of Collin Giguere et al 45 to generate bulk tissue genomic sequencing data from corresponding single-cell data, and subsequently, absolute copy numbers were extracted (Figure S8). Then we fitted the generated bulk tissue CNA profiles with the single-cell CNA signatures. Comparing the differences reveals some inconsistencies between the CNA signatures of bulk tissue and corresponding single cells (Figs. 5 A-B). This inconsistency may be attributed to the CNA masking effects of bulk tissue samples. In the analysis of CNA patterns in individual samples, such as sample P04, we noted that the scSig4 constituted 96.7% in bulk tissue, whereas in corresponding single-cell samples, most cells predominantly exhibited scSig3 (Figs. 5 C-D and Figure S9). These findings suggest that in highly heterogeneous cancers like HCC, CNA signatures extracted from bulk tissue may not accurately reflect the true nature of the actual CNA in each cancer cells. Single cell CNA signature analysis reveals the evolutionary trajectory of HCC CNA signatures We constructed a phylogenetic tree of CNA signatures to analyze the dynamic changes in the proportion of scSig across different single cells. The analysis revealed that scSig7 (characterized with small CNA) consistently appeared earlier in the evolutionary timeline. As the tumor progressed, the proportion of scSig1 and scSig2 (both are characterized with large CNA) gradually increased. These findings suggest that scSig7 happens in the early stages of tumorigenesis, while scSig1 and scSig2 are likely associated with the relatively late stages of HCC (Fig. 6 A and Figure S10). To further investigate the evolutional order of CNA signatures during HCC progression, we calculated the shared CNA breakpoints in individual cells to infer the potential evolutionary trajectories of tumor cells. The results indicated that scSig7 consistently appeared prior to other CNA signatures, while scSig1 and scSig2 were observed at later stages (Fig. 6 B and Figure S11). Additionally, we utilized the cancer CNA phylogenetic inference tool MEDICC2 46 to construct a CNA evolutionary tree. Based on the pseudotemporal ordering of each cell in the evolutionary tree, we analyzed the dynamic changes of scSigs. Consistent with the previous findings, with scSig7 appearing at an earlier stage (Figure S12). These results elucidate the evolutionary trajectory of tumor genomic CNA signature, demonstrating that amplifications of small DNA fragments occur before large-scale genomic alterations in HCC evolution. The accumulation of these small fragment changes may affect the progression of the cell cycle, leading to chromosomal-level or genome-wide DNA alterations 47 . Single-cell CNA signatures predict prognosis and drug response of HCC patients Bulk tissue CNA signatures has been reported to be associated with cancer clinical outcomes 6 , 7 , 11 , 48 . To compare the prognostic impact of single-cell CNA signatures with those derived from WGS and SNP array CNA signatures, we analyzed HCC samples with survival data from the PCAWG and TCGA cohorts. We calculated the activity of scSig, WGS_Sig, and SNP_Sig in these samples through signature filtering. Across datasets from different technological platforms, scSig show more robust performance compared with WGS_Sig, and SNP_Sig as reflected in the consistent prognostic outcomes in different datasets (Fig. 7 A). In two independent datasets, scSig1 and scSig4 were associated with better prognosis, while scSig3 was related to poorer outcomes in the SNP dataset, with a similar trend observed in the WGS dataset (Fig. 7 B and Figure S13A-B). This finding suggests the superiority of single-cell CNA signatures over traditional bulk tissue CNA signatures in terms of HCC prognosis prediction. We analyzed the correlations between drug sensitivity and CNA signatures. Our findings indicate that the sensitivity of HCC patients to the commonly used drug sorafenib is negatively correlated with the scSig7. Additionally, certain drugs demonstrated different responses to particular CNA signatures. Inhibitors of protein arginine methyltransferase (PRMT) have been found to impair tumorigenesis in HCC mouse models 49 – 51 . GSK591, a PRMT5 inhibitor, is regulated by MYC in HCC and has been identified as a direct target gene of MYC 52 , 53 . Patients carrying the scSig1 signature exhibited heightened sensitivity to GSK591, whereas they were non-responsive or showed resistance to other drugs (Figure S13C). Importantly, the scSig3 pattern, indicative of normal genomic features, was sensitive to most drugs, suggesting that early-stage tumors without significant CNAs have not yet developed substantial drug resistance. As genomic CNAs increase, tumors gradually acquire drug resistance. These results provide a new perspective for using single-cell CNA signatures in studying tumor drug resistance. Discussion In this study, we have developed an analytical framework for the extraction of CNA signatures in single cells. This framework is adaptable across various experimental platforms and sequencing depths, effectively extracting CNA signatures. CNA signatures have demonstrated their clinical value in a multitude of cancer studies 11 , 15 . Nevertheless, these existing CNA signature analyzes are all performed using bulk tissue samples, and single-cell CNA signature analysis is needed. We have systematically compared CNA signatures in single cells with those in bulk tissues of HCC. This study demonstrates that the majority of CNA signatures identified at the single-cell level correspond closely with those found in bulk tissue sequencing data of HCC. By employing existing methods to simulate CNA mutational processes and CNA noise, this research evaluated the reliability of different CNA signatures. The findings indicate that single-cell CNA signatures reveal potential CNA mutational processes with higher precision and exhibit greater stability in the presence of simulated noise. Despite the high heterogeneity of HCC, single-cell data accurately reflects the true state of CNA, as it is not affected by the mixing of subclones with different CNA. This study is the first attempt to depict the dynamic changes of CNA signatures during HCC progression. We found that CNA signatures with small fragment amplifications, such as scSig7, typically appear in the early stages of HCC development. In contrast, CNA signatures indicative of chromosomal and genomic-level changes, such as scSig1 and scSig2, tend to emerge in the later stages. A reasonable explanation for this pattern is that amplification of small DNA fragments in the early stages disrupts normal cell cycle processes, subsequently leading to large-scale genomic variations, which contribute to the formation of complex large CNA patterns. However, these conclusions still require direct experimental evidence for further validation. CNA is a complex biological process, and studies focusing on bulk tissues have already demonstrated its potential clinical utility. Nevertheless, the biological significance of certain CNA patterns remains to be elucidated, necessitating further in-depth investigation. In this study, we show that CNA signatures derived from single cells can effectively predict patient prognosis and drug sensitivity. Combining single-cell data with bulk tissue CNA data for integrated analysis could be a promising avenue for future research. From a single-cell perspective, we have identified some potential issues with bulk tissue CNA signatures. However, our study also has certain limitations. Although this study included data from 1,222 single cells, these data were derived from only 10 patients. The limited sample size might have led to the failure to detect all CNA features. Regarding the evolutionary trajectory analysis of CNA signatures, our analysis suggest that certain signatures consistently appear earlier in different samples. However, the complexity of the CNA process makes it challenging to establish an accurate evolutionary timeline and comprehensively describe the detailed evolutionary trajectory of CNA signatures. These analyses provide a direction for future research, and more experimental evidence is required for validation. Resource availability Lead contact Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Xue-Song Liu ( [email protected] ) Materials availability This study did not generate new materials. Data and code availability Only publicly available data were used in this study, and data sources and handling of these data are described in the Materials and Methods. All codes required to reproduce the results reported in this manuscript are freely available at: https://github.com/XSLiuLab/single-cell-CNA-signature . Further information is available from the corresponding author upon request. Method details Data collection and processing We downloaded data from 1222 single-cell shallow whole-genome sequencing (0.4×) of HCC patient samples (totaling 10 patients) from the Genome Sequence Archive (GSA) 24 , 27 , under accession numbers HRA000094. Data from 160 HCC bulk samples with shallow whole-genome sequencing (0.5×) were also obtained from the International Cancer Genome Consortium (ICGC) database 28 . Moreover, absolute copy number data of whole-genome sequencing from 178 HCC bulk samples in the Pan-Cancer Analysis of Whole Genomes (PCAWG) cohort were acquired through the UCSC Xena platform 29 , 30 . Concurrently, SNP array data from 371 HCC bulk samples were sourced from The Cancer Genome Atlas (TCGA) database 31 . For genome sequencing raw data (fastq) files, fastp was first used to perform quality control on the data, the BWA MEM algorithm was then applied using hg38 as reference genome 54 , and finally, the samtools tool was used to convert the SAM files into BAM files 55 . After obtaining the BAM files for all datasets, use the SortSam and MarkDuplicates commands in the Picard toolkit ( https://broadinstitute.github.io/picard/ ) to sort BAM files and mark duplications for variant and copy number calling. Raw data processing the codes used can be found at Code availability. Absolute copy number calling QDNAseq 56 and ACE 57 R packages were used to determine the copy number profiles from the low-coverage WGS BAM files. To determine the autosomal copy number profiles, we excluded both chromosome X&Y and mitochondrial DNA and the ploidy was adjusted using median bin segment value, which was the central assumption of ACE. For single-cell sequencing files, we used a ploidy penalty of 0.5 and lower-cellularity penalty of 1.0 to fit the “squaremodel()” function of ACE. For bulk tissue sequencing files, we used a ploidy penalty of 0.5 and lower-cellularity penalty of 0.5 to fit the “squaremodel()” function of ACE as per the author’s recommendation. ABSOLUTE 58 was used to determine the copy number profiles from the SNP array files of TCGA. The code for copy number calling can be found in Code availability. CNA feature classification Considering that shallow sequencing data cannot assess the LOH status of the genome, in order to describe the copy number variation of the liver cancer genome more accurately, we classify the copy number segments from four aspects, (1): copy number, which directly reflects the copy number amplification or deletion, divided into 6 states: 0, 1, 2, 3, 4, 5+, where 0 and 1 indicate that the genome fragment is missing, 2 indicates the normal genome state, and 3 and 4 indicate that the corresponding genome fragment is moderately amplified., 5 + is considered highly amplified. (2): segment length, indicating the length of the copy number segment. The segment length of some small genomic variations such as ecDNA is mostly less than 5Mb 59 , 60 . Since most of the copy number variations produced by large scale transitions are greater than 10MB 61 , the copy number segment is length is divided into three types: S (5Mb 10Mb). (3): segment change, indicating the change in copy number between two segments. Different mutations may imply that the mutations in adjacent segments come from different mechanisms. Divided into two types: AA ( 2). (4): segment shape, the relationship between the copy number segment and two adjacent segments, divided into three types: HH (left High and right High), LL (left Low and right Low), and OT (other status). Deriving signatures Copy number signatures were extracted from individual sample matrices using the gold standard tool SigProfiler v1.0.17 ( https://github.com/AlexandrovLab/SigProfilerExtractor ) with default parameters. Initially, the non-negative matrix factorization (NMF) algorithm was applied for the decomposition of the patient's copy number feature matrix. Signature search intervals ranging from 3 to 12 were set, and NMF was run 200 times for each signature number with different random seeds to ensure result stability. For each NMF iteration, initialization was performed using random numbers, and iterations ranged from 10,000 to 1,000,000 until stable results were obtained. Two key parameters determined during this process were the stability of the signatures, as measured by silhouette analysis, and the average Frobenius reconstruction error. Under these criteria, seven signatures were identified as optimal at the single-cell level and were selected for further analysis. Copy number variation simulation To simulate realistic chromosomal variation events, this study utilized the approach proposed by Drews et al 7 . We simulated WGD events by setting a tetraploid background prior to the insertion of other CNAs for early WGD, or by multiplying all copy number states by a factor following the placement of all CNAs for late WGD. Considering that most WGD events are early occurrences, the ratio of early to late WGD samples was set at 4:1 62 . The simulation of LST was based on their real breakpoint characteristics, employing a Poisson distribution for this simulation. Random copy number variations were simulated using the RSVSim 44 tool, with hg38 as the reference genome. The simulations aimed to cover 0–15% of the genome. The length of each amplified and deleted segment was randomly set between 1kb and 5MB, distributed randomly across the entire genome, with all other parameters set to their default values. Calculating the degree of similarity between simulated samples and scSig using cosine similarity, a method widely used for assessing similarities among copy number profiles 8 . Phylogenic trees We utilized the phangorn package in R to calculate the Maximum Parsimony Tree from the scSig matrix using the Parsimony Ratchet algorithm 63 , 64 . To ensure the accuracy of our analysis, samples that did not exhibit any CNAs were excluded. The branch lengths between cells were estimated using the ACCTRAN algorithm. The "ggtree" package in R was employed to visualize the phylogenetic trees 65 , where the predominant CNA signature were marked with distinct colors for easy visual differentiation. CNA signature evolution To determine the temporal sequence of emergence for tumor cell subclones, we estimated the likely order by analyzing shared CNA breakpoints. Shared breakpoints refer to genomic CNAs that occur at the same positions across different subclonal cells within the same sample. Each subclone’s CNA may build upon the segments from prior CNA events, but these subclones are likely to retain the breakpoints from earlier CNAs. where S j is the frequency of shared break point of cell j in the sample, M is the number of cells in the sample, L j is the number of shared breakpoints in cell j, and I ij represents the existence of the shared breakpoint I in cell j. I ij is a binary variable. When the breakpoint i exists in the cell j, I ij takes a value of 1; when the breakpoint i does not exist in the cell j, I ij takes a value of 0. Survival analysis In this study, we gathered survival data for samples from the TCGA (SNP array sequencing) and PCAWG (WGS sequencing) projects, encompassing 352 samples from TCGA and 139 samples from PCAWG. R package sigminer 15 was utilized to fit each patient's copy number feature matrix to scSig, SNP_Sig, and WGS_Sig, respectively. The association between CNA feature activity and overall survival was determined using univariate Cox proportional hazards models. For each Cox model, Hazard ratio were calculated and reported to indicate the direction and significance of the survival relationship. The Cox model is given by: where t is the total survival time, h(t, X) is the hazard function given the presence of a prognostic variable X, and h0(t) represents the baseline hazard. Cox proportional hazards analysis was conducted using the R package ezcox 66 . Drug sensitivity analysis Genomic information and drug sensitivity data for all HCC cell lines were downloaded from the GDSC database 67 . Following the previously introduced method, copy number variation data were obtained, and copy number feature matrices were calculated. Subsequently, using the R package sigminer, the copy number feature matrices of each cell line were fitted to scSig. Spearman correlation analysis was utilized to assess the relationship between the exposure values of scSig for each sample and IC50. The significance P-values obtained were corrected using the FDR method. Finally, the R package ComplexHeatmap 68 was employed to draw correlation heatmap. Declarations Acknowledgments We thank ShanghaiTech University High Performance Computing Public Service Platform for computing services. We thank multi-omics facility, molecular and cell biology core facility of ShanghaiTech University for technical help. This work is supported by cross disciplinary Research Fund of Shanghai Ninth People's Hospital, Shanghai JiaoTong University School of Medicine (JYJC202227). Shanghai Science and Technology Commission (21ZR1442400), National Natural Science Foundation of China (82373149), and startup funding from ShanghaiTech University. Author Contributions C.W. performed all the CNA signature analysis and drafted the manuscript under the supervision of X.-S. L. Z.T., W.C., N.W., J.W., J.S., T.W., D.Q., K.D., X.Z. T.L., L.Z., W.W., X.L., X.L., participated in cancer genome analysis and discussion. X.X. provided critical materials for this study and participated in critical project discussion and supervision. X.-S. L. conceived this study, supervised this study and wrote the manuscript with C.W. Declaration of interests We declare no competing interest. References Hanahan, D. (2022). Hallmarks of Cancer: New Dimensions. Cancer discovery 12 , 31-46. 10.1158/2159-8290.Cd-21-1059. Alexandrov, L.B., Nik-Zainal, S., Wedge, D.C., Aparicio, S.A.J.R., Behjati, S., Biankin, A.V., Bignell, G.R., Bolli, N., Borg, A., Børresen-Dale, A.-L., et al. (2013). Signatures of mutational processes in human cancer. Nature 500 , 415-421. 10.1038/nature12477. Poulsgaard, G.A., Sørensen, S.G., Juul, R.I., Nielsen, M.M., and Pedersen, J.S. (2023). Sequence dependencies and mutation rates of localized mutational processes in cancer. Genome Medicine 15 , 63. 10.1186/s13073-023-01217-z. Yang, G., Shahatiaili, A., Bai, S., Wang, L., Jin, D., Cao, M., Su, P., Liu, Q., Tao, K., and Long, Q.J.T.J.o.P. (2024). Mutational signature and prognosis in adenocarcinoma of the bladder. 262 , 334-346. Alexandrov, L.B., Kim, J., Haradhvala, N.J., Huang, M.N., Tian Ng, A.W., Wu, Y., Boot, A., Covington, K.R., Gordenin, D.A., Bergstrom, E.N., et al. (2020). The repertoire of mutational signatures in human cancer. Nature 578 , 94-101. 10.1038/s41586-020-1943-3. Steele, C.D., Abbasi, A., Islam, S.A., Bowes, A.L., Khandekar, A., Haase, K., Hames-Fathi, S., Ajayi, D., Verfaillie, A., and Dhami, P.J.N. (2022). Signatures of copy number alterations in human cancer. 606 , 984-991. Drews, R.M., Hernando, B., Tarabichi, M., Haase, K., Lesluyes, T., Smith, P.S., Morrill Gavarró, L., Couturier, D.-L., Liu, L., and Schneider, M.J.N. (2022). A pan-cancer compendium of chromosomal instability. 606 , 976-983. Tao, Z., Wang, S., Wu, C., Wu, T., Zhao, X., Ning, W., Wang, G., Wang, J., Chen, J., and Diao, K.J.B.i.B. (2023). The repertoire of copy number alteration signatures in human cancer. 24 , bbad053. Wang, S., Tao, Z., Wu, T., and Liu, X.-S.J.B. (2021). Sigflow: an automated and comprehensive pipeline for cancer genome mutational signature analysis. 37 , 1590-1592. Wang, S., Jia, M., He, Z., and Liu, X.-S.J.O. (2018). APOBEC3B and APOBEC mutational signature as potential predictive markers for immunotherapy response in non-small cell lung cancer. 37 , 3924-3936. Macintyre, G., Goranova, T.E., De Silva, D., Ennis, D., Piskorz, A.M., Eldridge, M., Sie, D., Lewsley, L.-A., Hanif, A., Wilson, C., et al. (2018). Copy number signatures and mutational processes in ovarian carcinoma. Nature Genetics 50 , 1262-1270. 10.1038/s41588-018-0179-8. Doran, C.G., and Pennington, S.R.J.B.i.M. (2022). Copy number alteration signatures as biomarkers in cancer: a review. 16 , 371-386. Essers, P.B., van der Heijden, M., Vossen, D., de Roest, R.H., Leemans, C.R., Brakenhoff, R.H., van den Brekel, M.W., Bartelink, H., Verheij, M., and Vens, C.J.I.j.o.c. (2020). Ovarian cancer‐derived copy number alterations signatures are prognostic in chemoradiotherapy‐treated head and neck squamous cell carcinoma. 147 , 1732-1739. Yao, H., Li, H., Wang, J., Wu, T., Ning, W., Diao, K., Wu, C., Wang, G., Tao, Z., and Zhao, X.J.C.B. (2023). Copy number alteration features in pan-cancer homologous recombination deficiency prediction and biology. 6 , 527. Wang, S., Li, H., Song, M., Tao, Z., Wu, T., He, Z., Zhao, X., Wu, K., and Liu, X.-S.J.P.G. (2021). Copy number signature analysis tool and its application in prostate cancer reveals distinct mutational processes and clinical outcomes. 17 , e1009557. Nguyen, P.H.D., Ma, S., Phua, C.Z.J., Kaya, N.A., Lai, H.L.H., Lim, C.J., Lim, J.Q., Wasser, M., Lai, L., Tam, W.L., et al. (2021). Intratumoural immune heterogeneity as a hallmark of tumour evolution and progression in hepatocellular carcinoma. Nature Communications 12 , 227. 10.1038/s41467-020-20171-7. Li, L., and Wang, H.J.C.l. (2016). Heterogeneity of liver cancer and personalized therapy. 379 , 191-197. Hiley, C., de Bruin, E.C., McGranahan, N., and Swanton, C.J.G.b. (2014). Deciphering intratumor heterogeneity and temporal acquisition of driver events to refine precision medicine. 15 , 1-10. Lei, H., Lyu, B., Gertz, E.M., Schäffer, A.A., Shi, X., Wu, K., Li, G., Xu, L., Hou, Y., and Dean, M.J.J.o.C.B. (2020). Tumor copy number deconvolution integrating bulk and single-cell sequencing data. 27 , 565-598. Malikic, S., Mehrabadi, F.R., Ciccolella, S., Rahman, M.K., Ricketts, C., Haghshenas, E., Seidman, D., Hach, F., Hajirasouliha, I., and Sahinalp, S.C. (2019). PhISCS: a combinatorial approach for subperfect tumor phylogeny reconstruction via integrative use of single-cell and bulk sequencing data. Genome research 29 , 1860-1877. 10.1101/gr.234435.118. Graham, T.A., and Sottoriva, A.J.T.J.o.p. (2017). Measuring cancer evolution from the genome. 241 , 183-191. Turajlic, S., Sottoriva, A., Graham, T., and Swanton, C.J.N.R.G. (2019). Resolving genetic heterogeneity in cancer. 20 , 404-416. Gao, R., Davis, A., McDonald, T.O., Sei, E., Shi, X., Wang, Y., Tsai, P.-C., Casasent, A., Waters, J., and Zhang, H.J.N.g. (2016). Punctuated copy number evolution and clonal stasis in triple-negative breast cancer. 48 , 1119-1130. Guo, L., Yi, X., Chen, L., Zhang, T., Guo, H., Chen, Z., Cheng, J., Cao, Q., Liu, H., Hou, C., et al. (2022). Single-Cell DNA Sequencing Reveals Punctuated and Gradual Clonal Evolution in Hepatocellular Carcinoma. Gastroenterology 162 , 238-252. 10.1053/j.gastro.2021.08.052. Gerstung, M., Jolly, C., Leshchiner, I., Dentro, S.C., Gonzalez, S., Rosebrock, D., Mitchell, T.J., Rubanova, Y., Anur, P., and Yu, K.J.N. (2020). The evolutionary history of 2,658 cancers. 578 , 122-128. McPherson, A., Roth, A., Laks, E., Masud, T., Bashashati, A., Zhang, A.W., Ha, G., Biele, J., Yap, D., and Wan, A.J.N.g. (2016). Divergent modes of clonal spread and intraperitoneal mixing in high-grade serous ovarian cancer. 48 , 758-767. Database Resources of the National Genomics Data Center, China National Center for Bioinformation in 2023. (2023). Nucleic acids research 51 , D18-d28. 10.1093/nar/gkac1073. Zhang, J., Bajari, R., Andric, D., Gerthoffert, F., Lepsa, A., Nahal-Bose, H., Stein, L.D., and Ferretti, V.J.N.b. (2019). The international cancer genome consortium data portal. 37 , 367-369. Pan-cancer analysis of whole genomes %J Nature. (2020). 578 , 82-93. Goldman, M.J., Craft, B., Hastie, M., Repečka, K., McDade, F., Kamath, A., Banerjee, A., Luo, Y., Rogers, D., Brooks, A.N., et al. (2020). Visualizing and interpreting cancer genomics data via the Xena platform. Nature Biotechnology 38 , 675-678. 10.1038/s41587-020-0546-8. Zack, T.I., Schumacher, S.E., Carter, S.L., Cherniack, A.D., Saksena, G., Tabak, B., Lawrence, M.S., Zhang, C.-Z., Wala, J., and Mermel, C.H.J.N.g. (2013). Pan-cancer patterns of somatic copy number alteration. 45 , 1134-1140. Korbel, J.O., and Campbell, P.J.J.C. (2013). Criteria for inference of chromothripsis in cancer genomes. 152 , 1226-1236. Popova, T., Manié, E., Rieunier, G., Caux-Moncoutier, V., Tirapo, C., Dubois, T., Delattre, O., Sigal-Zafrani, B., Bollet, M., and Longy, M.J.C.r. (2012). Ploidy and large-scale genomic instability consistently identify basal-like breast carcinomas with BRCA1/2 inactivation. 72 , 5454-5462. Al-Rawi, D.H., Bakhoum, S.F.J.C.O.i.G., and Development (2022). Chromosomal instability as a source of genomic plasticity. 74 , 101913. Menghi, F., Inaki, K., Woo, X., Kumar, P.A., Grzeda, K.R., Malhotra, A., Yadav, V., Kim, H., Marquez, E.J., and Ucar, D.J.P.o.t.N.A.o.S. (2016). The tandem duplicator phenotype as a distinct genomic configuration in cancer. 113 , E2373-E2382. Alexandrov, L.B., Nik-Zainal, S., Wedge, D.C., Aparicio, S.A., Behjati, S., Biankin, A.V., Bignell, G.R., Bolli, N., Borg, A., and Børresen-Dale, A.-L.J.N. (2013). Signatures of mutational processes in human cancer. 500 , 415-421. Alexandrov, L.B., Kim, J., Haradhvala, N.J., Huang, M.N., Tian Ng, A.W., Wu, Y., Boot, A., Covington, K.R., Gordenin, D.A., and Bergstrom, E.N.J.N. (2020). The repertoire of mutational signatures in human cancer. 578 , 94-101. Kim, H., Nguyen, N.-P., Turner, K., Wu, S., Gujar, A.D., Luebeck, J., Liu, J., Deshpande, V., Rajkumar, U., and Namburi, S.J.N.g. (2020). Extrachromosomal DNA is associated with oncogene amplification and poor outcome across multiple cancers. 52 , 891-897. Yi, E., Chamorro González, R., Henssen, A.G., and Verhaak, R.G.W. (2022). Extrachromosomal DNA amplifications in cancer. Nature Reviews Genetics 23 , 760-771. 10.1038/s41576-022-00521-5. Zhao, X.-K., Xing, P., Song, X., Zhao, M., Zhao, L., Dang, Y., Lei, L.-L., Xu, R.-H., Han, W.-L., and Wang, P.-P.J.N.c. (2021). Focal amplifications are associated with chromothripsis events and diverse prognoses in gastric cardia adenocarcinoma. 12 , 6489. Koche, R.P., Rodriguez-Fos, E., Helmsauer, K., Burkert, M., MacArthur, I.C., Maag, J., Chamorro, R., Munoz-Perez, N., Puiggros, M., and Dorado Garcia, H.J.N.g. (2020). Extrachromosomal circular DNA drives oncogenic genome remodeling in neuroblastoma. 52 , 29-34. Dentro, S.C., Leshchiner, I., Haase, K., Tarabichi, M., Wintersinger, J., Deshwar, A.G., Yu, K., Rubanova, Y., Macintyre, G., and Demeulemeester, J.J.C. (2021). Characterizing genetic intra-tumor heterogeneity across 2,658 human cancer genomes. 184 , 2239-2254. e2239. Davoli, T., de Lange, T.J.A.r.o.c., and biology, d. (2011). The causes and consequences of polyploidy in normal development and cancer. 27 , 585-610. Bartenhagen, C., and Dugas, M.J.B. (2013). RSVSim: an R/Bioconductor package for the simulation of structural variations. 29 , 1679-1681. Giguere, C., Dubey, H.V., Sarsani, V.K., Saddiki, H., He, S., and Flaherty, P.J.B.b. (2020). SCSIM: Jointly simulating correlated single-cell and bulk next-generation DNA sequencing data. 21 , 1-10. Kaufmann, T.L., Petkovic, M., Watkins, T.B., Colliver, E.C., Laskina, S., Thapa, N., Minussi, D.C., Navin, N., Swanton, C., and Van Loo, P. (2022). MEDICC2: whole-genome doubling aware copy-number phylogenies for cancer evolution. Genome biology 23 , 241. Comaills, V., and Castellano-Pozo, M.J.B. (2023). Chromosomal instability in genome evolution: from cancer to macroevolution. 12 , 671. Maclachlan, K.H., Rustad, E.H., Derkach, A., Zheng-Lin, B., Yellapantula, V., Diamond, B., Hultcrantz, M., Ziccheddu, B., Boyle, E.M., and Blaney, P.J.N.c. (2021). Copy number signatures predict chromothripsis and clinical outcomes in newly diagnosed multiple myeloma. 12 , 5172. Lei, Y., Han, P., and Tian, D. (2021). Protein arginine methyltransferases and hepatocellular carcinoma: A review. Translational Oncology 14 , 101194. Gou, Q., He, S., and Zhou, Z. (2017). Protein arginine N-methyltransferase 1 promotes the proliferation and metastasis of hepatocellular carcinoma cells. Tumor Biology 39 , 1010428317691419. Jeon, J.-Y., Lee, J.S., Park, E.-R., Shen, Y.N., Kim, M.-Y., Shin, H.-J., Joo, H.-Y., Cho, E.-H., Moon, S.M., and Shin, U.S. (2018). Protein arginine methyltransferase 5 is implicated in the aggressiveness of human hepatocellular carcinoma and controls the invasive activity of cancer cells. Oncology reports 40 , 536-544. Luo, Y., Gao, Y., Liu, W., Yang, Y., Jiang, J., Wang, Y., Tang, W., Yang, S., Sun, L., and Cai, J. (2021). Myelocytomatosis‐protein arginine N‐methyltransferase 5 Axis defines the tumorigenesis and immune response in hepatocellular carcinoma. Hepatology 74 , 1932-1951. Dang, C.V. (2012). MYC on the path to cancer. Cell 149 , 22-35. Li, H.J.a.p.a. (2013). Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G., Abecasis, G., Durbin, R., and bioinformatics, G.P.D.P.S.J. (2009). The sequence alignment/map format and SAMtools. 25 , 2078-2079. Scheinin, I., Sie, D., Bengtsson, H., Van De Wiel, M.A., Olshen, A.B., Van Thuijl, H.F., Van Essen, H.F., Eijk, P.P., Rustenburg, F., and Meijer, G.A.J.G.r. (2014). DNA copy number analysis of fresh and formalin-fixed specimens by shallow whole-genome sequencing with identification and exclusion of problematic regions in the genome assembly. 24 , 2022-2032. Poell, J.B., Mendeville, M., Sie, D., Brink, A., Brakenhoff, R.H., and Ylstra, B.J.B. (2019). ACE: absolute copy number estimation from low-coverage whole-genome sequencing data. 35 , 2847-2849. Carter, S.L., Cibulskis, K., Helman, E., McKenna, A., Shen, H., Zack, T., Laird, P.W., Onofrio, R.C., Winckler, W., and Weir, B.A.J.N.b. (2012). Absolute quantification of somatic DNA alterations in human cancer. 30 , 413-421. Wu, S., Turner, K.M., Nguyen, N., Raviram, R., Erb, M., Santini, J., Luebeck, J., Rajkumar, U., Diao, Y., and Li, B.J.N. (2019). Circular ecDNA promotes accessible chromatin and high oncogene expression. 575 , 699-703. Wu, T., Wu, C., Zhao, X., Wang, G., Ning, W., Tao, Z., Chen, F., and Liu, X.-S.J.S.r. (2022). Extrachromosomal DNA formation enables tumor immune escape potentially through regulating antigen presentation gene expression. 12 , 3590. Schonhoft, J.D., Zhao, J.L., Jendrisak, A., Carbone, E.A., Barnett, E.S., Hullings, M.A., Gill, A., Sutton, R., Lee, J., and Dago, A.E.J.C.r. (2020). Morphology-predicted large-scale transition number in circulating tumor cells identifies a chromosomal instability biomarker associated with poor outcome in castration-resistant prostate cancer. 80 , 4892-4903. Bielski, C.M., Zehir, A., Penson, A.V., Donoghue, M.T., Chatila, W., Armenia, J., Chang, M.T., Schram, A.M., Jonsson, P., and Bandlamudi, C.J.N.g. (2018). Genome doubling shapes the evolution and prognosis of advanced cancers. 50 , 1189-1195. Paradis, E., and Schliep, K.J.B. (2019). ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R. 35 , 526-528. Schliep, K.P.J.B. (2011). phangorn: phylogenetic analysis in R. 27 , 592-593. Yu, G., Lam, T.T.-Y., Zhu, H., Guan, Y.J.M.b., and evolution (2018). Two methods for mapping and visualizing associated data on phylogeny using ggtree. 35 , 3041-3043. Wang, S., Liu, X.-S., Li, J., and Zhao, Q.J.a.p.a. (2021). Ezcox: an R/CRAN package for cox model batch processing and visualization. Yang, W., Soares, J., Greninger, P., Edelman, E.J., Lightfoot, H., Forbes, S., Bindal, N., Beare, D., Smith, J.A., and Thompson, I.R.J.N.a.r. (2012). Genomics of Drug Sensitivity in Cancer (GDSC): a resource for therapeutic biomarker discovery in cancer cells. 41 , D955-D961. Gu, Z., and Hübschmann, D.J.B. (2022). Make interactive complex heatmaps in R. 38 , 1460-1462. Additional Declarations There is NO Competing Interest. Supplementary Files suppFigure.pdf Supplemnet figure Cite Share Download PDF Status: Published Journal Publication published 17 Nov, 2025 Read the published version in Communications Biology → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6304422","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":437900442,"identity":"c472649b-333c-45a7-8f86-51c3b1d67a3a","order_by":0,"name":"Xue-Song Liu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABD0lEQVRIiWNgGAWjYBACPmYQWcGQACTZGBgMgJQEEPPg0cIG1nIGrsWACC0ggrENpoWBGC3sPGaSP+fV5fHPbr/24EfBHzn+2Q2MD962Mcib43QYj7Ex77bDxRJ3zpQb9hgYGEvcOcBsOLeNwXBnA04tho8Ztx1IbLiRkybBY2CQuEEigU2aF+hUgwM4tRgc/DmnLnE+UIvkHwODeqAW9t8EtBg+4G1gTtxwI/2YNNCWBAOgLcz4tbAVG/McO5y48UYOm7SMgbHhjBuJzZJzzkkYbsChhZ//8DbJHzV1ifNupD+TfPNHTp5/RvLBD2/KbORx2YIEeAygDMYGBkjsEATsD4hRNQpGwSgYBSMQAAC5BFNS1giT3wAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-7736-0077","institution":"ShanghaiTech University","correspondingAuthor":true,"prefix":"","firstName":"Xue-Song","middleName":"","lastName":"Liu","suffix":""},{"id":437900443,"identity":"1da561ea-ee89-493c-891a-747b2aede134","order_by":1,"name":"Chenxu Wu","email":"","orcid":"","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Chenxu","middleName":"","lastName":"Wu","suffix":""},{"id":437900444,"identity":"4bdd2daa-14fd-431d-948f-715c2fc09d72","order_by":2,"name":"Ziyu Tao","email":"","orcid":"https://orcid.org/0000-0003-3272-1227","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Ziyu","middleName":"","lastName":"Tao","suffix":""},{"id":437900445,"identity":"78320e0e-a3a2-4d63-9270-500e56be3aaf","order_by":3,"name":"Wen Chen","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Wen","middleName":"","lastName":"Chen","suffix":""},{"id":437900446,"identity":"6434bd0a-3549-4718-ba79-101fcfb57cce","order_by":4,"name":"Nan Wang","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Nan","middleName":"","lastName":"Wang","suffix":""},{"id":437900447,"identity":"928867f3-b5ad-43ce-9b8d-4d1066e7fd40","order_by":5,"name":"Jinyu Wang","email":"","orcid":"https://orcid.org/0000-0001-9033-4589","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Jinyu","middleName":"","lastName":"Wang","suffix":""},{"id":437900448,"identity":"62a2c743-ce84-4a87-9c1c-0bf6e094c1d5","order_by":6,"name":"Jiayu Shen","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Jiayu","middleName":"","lastName":"Shen","suffix":""},{"id":437900449,"identity":"e35bc765-cbc4-4750-b6ac-ce48a72083f7","order_by":7,"name":"Tao Wu","email":"","orcid":"https://orcid.org/0000-0002-8999-9628","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Tao","middleName":"","lastName":"Wu","suffix":""},{"id":437900450,"identity":"f6ab4e18-b108-4ee8-8961-5855bacb7bff","order_by":8,"name":"Die Qiu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Die","middleName":"","lastName":"Qiu","suffix":""},{"id":437900451,"identity":"86a7cfd3-4dcb-4caf-b41a-19d5cfe71318","order_by":9,"name":"Kaixuan Diao","email":"","orcid":"","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Kaixuan","middleName":"","lastName":"Diao","suffix":""},{"id":437900452,"identity":"0afba781-5484-411a-88f0-46892e649dc4","order_by":10,"name":"Xiangyu Zhao","email":"","orcid":"","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Xiangyu","middleName":"","lastName":"Zhao","suffix":""},{"id":437900453,"identity":"224babb3-9410-4327-9dd0-9f02a85817ef","order_by":11,"name":"Tianzhu Lu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Tianzhu","middleName":"","lastName":"Lu","suffix":""},{"id":437900454,"identity":"db508b7a-e1bf-4248-8ca7-133caaef4d55","order_by":12,"name":"Lin Zhang","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Lin","middleName":"","lastName":"Zhang","suffix":""},{"id":437900455,"identity":"870f42f4-a504-4336-9a79-29e29406ad7d","order_by":13,"name":"Weiliang Wang","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Weiliang","middleName":"","lastName":"Wang","suffix":""},{"id":437900456,"identity":"286522b4-da15-4bb1-8cf0-c88e588ccf51","order_by":14,"name":"Xinxing Li","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Xinxing","middleName":"","lastName":"Li","suffix":""},{"id":437900457,"identity":"c12ef14a-24c1-4bbf-b6ec-fb2d24593ca9","order_by":15,"name":"Xinxiang Li","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Xinxiang","middleName":"","lastName":"Li","suffix":""},{"id":437900458,"identity":"a39f83db-1799-44dd-9616-84d6c3af0275","order_by":16,"name":"Xiaopeng Xiong","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Xiaopeng","middleName":"","lastName":"Xiong","suffix":""}],"badges":[],"createdAt":"2025-03-25 13:50:58","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6304422/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6304422/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s42003-025-08994-w","type":"published","date":"2025-11-17T05:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":81281014,"identity":"24defc60-abae-41a5-8a4e-0095c899307f","added_by":"auto","created_at":"2025-04-24 10:08:44","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":255448,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6304422/v1/21e5f5fb5fa9baf8b1e387a7.png"},{"id":81281735,"identity":"f092ee1b-271f-43a5-8dc7-4e231e30c3b8","added_by":"auto","created_at":"2025-04-24 10:16:45","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":379850,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6304422/v1/2b64e17e42e1a5996d501c60.png"},{"id":81281061,"identity":"12d485d6-fc4e-4995-a817-b1332cdc59ba","added_by":"auto","created_at":"2025-04-24 10:08:47","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":607907,"visible":true,"origin":"","legend":"\u003cp\u003eProfile and potential mutational processes of single-cell CNA signature (scSig).\u003c/p\u003e\n\u003cp\u003eA, The CNA feature distribution of scSig, with the x-axis representing 90 types of CNA features and the y-axis indicating the contribution of each feature.\u003c/p\u003e\n\u003cp\u003eB, Representative CNA profile, prominent features and potential mechanisms for each\u003c/p\u003e\n\u003cp\u003eidentified scSig.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-6304422/v1/bcfebcb3b0c3eb08fb4c437a.png"},{"id":81281056,"identity":"5bddc9ba-f135-43c6-966f-9db23505763a","added_by":"auto","created_at":"2025-04-24 10:08:46","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":505731,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-6304422/v1/81284b5a57d468669136dbf0.png"},{"id":81281051,"identity":"14816b64-f851-47e8-a2ac-6ba1752a78c6","added_by":"auto","created_at":"2025-04-24 10:08:46","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":786076,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-6304422/v1/867da66b3c32bc66a60a2a5f.png"},{"id":81280989,"identity":"cbbcb902-0daf-46fa-916f-4022906cd23d","added_by":"auto","created_at":"2025-04-24 10:08:42","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":518079,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-6304422/v1/2d77beddbda4acf34204b518.png"},{"id":81281049,"identity":"56b32a51-b0d4-48ad-9b25-2e53cccf4dde","added_by":"auto","created_at":"2025-04-24 10:08:45","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":637567,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-6304422/v1/59453f42b8575ec4839a3344.png"},{"id":96153824,"identity":"5349be19-b09c-487c-9448-365359058816","added_by":"auto","created_at":"2025-11-18 08:07:36","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3834842,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6304422/v1/32367ac9-9000-4d8b-be27-4137a469d750.pdf"},{"id":81281065,"identity":"0b6c7e0c-bbaf-412e-b5f8-e603e3fb9796","added_by":"auto","created_at":"2025-04-24 10:08:48","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":11462212,"visible":true,"origin":"","legend":"Supplemnet figure","description":"","filename":"suppFigure.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6304422/v1/ed36280b5a13615341fcc8e4.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Single-cell copy number alteration signature analysis reveals masked patterns and potential biomarkers for cancer","fulltext":[{"header":"Highlights","content":"\u003cp\u003eSingle-cell analysis reveals CNA signatures masked in bulk tissue.\u003c/p\u003e\n\u003cp\u003eSingle-cell analysis delineates the evolutionary trajectory of CNA signature.\u003c/p\u003e\n\u003cp\u003eSmall CNAs occur early and large CNAs happens late in HCC evolution.\u003c/p\u003e\n\u003cp\u003eSingle-cell CNA signatures show robust performance in guiding cancer clinical treatment.\u003c/p\u003e"},{"header":"Introduction","content":"\u003cp\u003eGenomic instability is one of the hallmarks of cancer\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e, encompassing changes ranging from single nucleotide alterations to whole chromosome modifications. Numerous studies have demonstrated the roles of single base substitutions (SBS), small insertions and deletions (INDEL), structural variations (including translocations/inversions), and copy number alterations in the process of cancer development \u003csup\u003e\u003cspan additionalcitationids=\"CR3\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. Genomic DNA alteration signatures are recurring genomic patterns that are the imprints of mutagenic processes accumulated over the lifetime of cancer cell\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. SBS signature analysis has been extensively studied and represents a prototype for other types of signature study\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. While analyzing copy number signature is more intricate than SBS signatures. Pan-cancer CNA signature studies have been carried out\u003csup\u003e\u003cspan additionalcitationids=\"CR7 CR8\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e, and these studies indicate that genomic alteration signatures can provide information on mutational processes, and also serve as biomarkers for precision medicine in cancer\u003csup\u003e\u003cspan additionalcitationids=\"CR11 CR12 CR13 CR14\" citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eCurrent research on CNA signature is predominantly focused on bulk tissue samples. In the detection of SBS, bulk tissue is capable of preserving the SBS mutation status of the majority of subclones. In contrast, CNA usually affect large proportion of genome DNA, and the final CNA profile of bulk tissue is usually not the linear combination of each subclone's CNA state (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e as example). For tumors with low heterogeneity, sequencing of bulk tissue effectively captures the variations in copy numbers. However, for tumors with high heterogeneity, like hepatocellular carcinoma (HCC)\u003csup\u003e\u003cspan additionalcitationids=\"CR17\" citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e, sequencing of bulk tissue only reflects the averaged CNA across all tumor subclones\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e,\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e, failing to accurately reveal the true patterns of CNA.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe models of tumor evolution have been a contentious subject in tumor biology\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. Gao et al's model of punctuated copy number evolution suggests that copy number aberrations are primarily acquired during brief peaks of genomic crisis, followed by stable clonal expansion\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. Guo et al\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e introduced a model of dual-phase copy number evolution in HCC, indicating the coexistence of gradual and punctuated evolution. Although these tumor evolution models have revealed the occurrence rate of CNAs during tumor progression, they have not clarified which CNA patterns or signatures appear in the early stages of a tumor and which appear in the later stages. The evolutionary trajectory of CNA signatures still remain elusive.\u003c/p\u003e \u003cp\u003eIn this study, we collected scDNA-seq data from 1222 cells of 10 HCC patients and developed a versatile tool for analyzing CNA signatures that apply to both single-cell and bulk tissue samples. This tool enabled us to systematically analyze the CNA signatures in HCC at the single-cell level for the first time. Comparing CNA signatures derived from single-cell and those from bulk samples revealed that some CNA signatures observed in bulk tissue samples might not actually exist. Additionally, we have for the first time proposed the evolutionary trajectory of CNA signatures in HCC. These studies provide new insights into the mutation processes of HCC and important information for the in-depth study of CNA signatures and their potential clinical applications.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eSingle-cell analysis revealed masked CNA profiles compared with bulk tissue CNA analysis\u003c/h2\u003e \u003cp\u003eEach point mutation occurs on a single chromosome in a single cell, which gives rise to a lineage of cells bearing the same mutation. If that chromosomal locus is subsequently duplicated, any point mutation on this allele preceding the gain will subsequently be present on the two resulting allelic copies\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e,\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. This process facilitates the accumulation of SBS without interference between mutations. Bulk tissue sequencing enables point mutation detection within a sample, reflecting the cumulative outcome of all subclones (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). However CNA usually affect large proportion of genome DNA, when a CNA occurs within a single cell, the subsequent subclonal CNA further modify the original CNA profile, distorting its characteristic signature. Consequently, the CNA observed in bulk tissue is not an accurate reflection of the true CNA occurs in each subclones (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). This phenomenon is particularly pronounced in tumors with high heterogeneity, such as HCC. To accurately explore the true state of CNA, we aggregated data from various HCC databases, covering different sequencing platforms and depths. Our collection comprised 1222 single-cell shallow whole-genome sequencing (sWGS) samples\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e, 160 bulk sWGS samples\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e, 178 bulk whole-genome sequencing (WGS) samples\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e,\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e, and 371 bulk SNP array samples\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. All raw data were uniformly processed to derive the absolute copy number information of the samples (details provided in the Materials and Methods).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eDevelopment of CNA signature analysis tool for both single-cell and bulk tissue samples\u003c/h3\u003e\n\u003cp\u003eTo identify single-cell CNA signatures, we developed a novel method for the extraction of CNA signatures. This method encompasses four principal aspects of CNA: absolute copy number, segment length, segment change, and segment shape (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). From these aspects, we further delineated 90 distinct features (Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). These features were selected as hallmarks of previously reported genomic aberrations, including chromothripsis\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e, large-scale state transitions (LST)\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e, extrachromosomal circular DNA (ecDNA)\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e, and tandem duplications\u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. Following the computation of features for all samples, this feature matrix was processed using non-negative matrix factorization to identify CNA signatures (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). This method has previously been employed for extracting signatures of SBS\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e,\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. The number of signatures extracted was determined using two parameters. First is the reconstruction error and the average Frobenius reconstruction error is reported. Second is the stability of signature extraction and the cosine similarity between the extracted signatures, and average silhouette width. In the single-cell samples, a total of seven CNA signatures were identified. In the analysis of bulk samples, seven CNA signatures were identified with sWGS and WGS data. Eight CNA signatures were determined using SNP array data (Figure S2).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eMutational processes for single-cell CNA signatures\u003c/h3\u003e\n\u003cp\u003eFollowing determining CNA signatures, we initially assessed the similarity among different CNA signatures. The observed similarities between CNA signatures were low (Figure S3), suggesting that each CNA signature is relatively distinct (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e3\u003c/span\u003eA and Figure S4). Further, we compared the CNA signatures extracted from bulk tissue sWGS, WGS, and SNP array data, with those obtained from single-cell sWGS data. Our results indicate that the CNA signatures identified in single-cell shallow sequencing data can also be detected in other data types (Figure S5A-C), thus demonstrating the stability and reliability of our proposed CNA signature extraction method. However, specific CNA signatures, such as WGS_Sig5, WGS_Sig6, SNP_Sig3, and SNP_Sig4, did not have similar counterparts in single-cell CNA signatures. This phenomenon may be attributed to several factors. On the one hand, the limited number of patients represented in the single-cell samples may have resulted in the exclusion of certain sample-specific CNA signatures from the analysis. On the other hand, these CNA signatures may represent signals specific to bulk tissues rather than single cells.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eAssociation analysis between CNA signatures and phenotypic data can reveal the underlying clinical and biological information for CNA signatures\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e,\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. In the scSig1, we identified numerous genomic fragments smaller than 5Mb with high copy numbers, consistent with genomic variations induced by ecDNA\u003csup\u003e\u003cspan additionalcitationids=\"CR39 CR40\" citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. Further associative analyses suggest these variations are attributable to ecDNA (Figure S6). Based on the analysis of representative samples of CNA signatures, we observed that the scSig3 pattern reflects characteristics of a normal genome, while the genome with the scSig4 pattern exhibits a clear triploid state, aligning with the CNS3 and CNS12 proposed by Tao et al\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. Utilizing these analyses, we have summarized the potential etiologies of each CNA signature in single-cell data from HCC (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e3\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eSingle-cell CNA signature analysis reveals masked patterns in bulk tissue analysis\u003c/h3\u003e\n\u003cp\u003eTo further investigate the differences between single-cell CNA signatures and bulk tissue CNA signatures, following the methodology of Drews et al\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e, this study simulated CNA events, including early and late whole-genome duplication (early WGD, late WGD)\u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e,\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e and LST (associated with homologous recombination deficiency, HRD)\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. The results indicate that single-cell CNA signatures can distinguish early WGD from late WGD, however bulk tissue CNA signatures fail to effectively differentiate early from late WGD events (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). This finding reflects the advantag e of single-cell CNA signatures in finely distinguishing patterns of CNA. Following the genomic variation simulation method of Bartenhagen et al\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e, we simulated the formation of a series of CNA for comparative analysis, and the results also demonstrated that single-cell CNA signatures are more robust than bulk tissue CNA signatures in filtering out such simulated noise (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e4\u003c/span\u003eB and Figure S7).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWe employed the method of Collin Giguere et al\u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e to generate bulk tissue genomic sequencing data from corresponding single-cell data, and subsequently, absolute copy numbers were extracted (Figure S8). Then we fitted the generated bulk tissue CNA profiles with the single-cell CNA signatures. Comparing the differences reveals some inconsistencies between the CNA signatures of bulk tissue and corresponding single cells (Figs.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e5\u003c/span\u003eA-B). This inconsistency may be attributed to the CNA masking effects of bulk tissue samples. In the analysis of CNA patterns in individual samples, such as sample P04, we noted that the scSig4 constituted 96.7% in bulk tissue, whereas in corresponding single-cell samples, most cells predominantly exhibited scSig3 (Figs.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e5\u003c/span\u003eC-D and Figure S9). These findings suggest that in highly heterogeneous cancers like HCC, CNA signatures extracted from bulk tissue may not accurately reflect the true nature of the actual CNA in each cancer cells.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eSingle cell CNA signature analysis reveals the evolutionary trajectory of HCC CNA signatures\u003c/h3\u003e\n\u003cp\u003eWe constructed a phylogenetic tree of CNA signatures to analyze the dynamic changes in the proportion of scSig across different single cells. The analysis revealed that scSig7 (characterized with small CNA) consistently appeared earlier in the evolutionary timeline. As the tumor progressed, the proportion of scSig1 and scSig2 (both are characterized with large CNA) gradually increased. These findings suggest that scSig7 happens in the early stages of tumorigenesis, while scSig1 and scSig2 are likely associated with the relatively late stages of HCC (Fig.\u0026nbsp;\u003cspan refid=\"Fig15\" class=\"InternalRef\"\u003e6\u003c/span\u003eA and Figure S10). To further investigate the evolutional order of CNA signatures during HCC progression, we calculated the shared CNA breakpoints in individual cells to infer the potential evolutionary trajectories of tumor cells. The results indicated that scSig7 consistently appeared prior to other CNA signatures, while scSig1 and scSig2 were observed at later stages (Fig.\u0026nbsp;\u003cspan refid=\"Fig15\" class=\"InternalRef\"\u003e6\u003c/span\u003eB and Figure S11). Additionally, we utilized the cancer CNA phylogenetic inference tool MEDICC2\u003csup\u003e46\u003c/sup\u003e to construct a CNA evolutionary tree. Based on the pseudotemporal ordering of each cell in the evolutionary tree, we analyzed the dynamic changes of scSigs. Consistent with the previous findings, with scSig7 appearing at an earlier stage (Figure S12). These results elucidate the evolutionary trajectory of tumor genomic CNA signature, demonstrating that amplifications of small DNA fragments occur before large-scale genomic alterations in HCC evolution. The accumulation of these small fragment changes may affect the progression of the cell cycle, leading to chromosomal-level or genome-wide DNA alterations\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eSingle-cell CNA signatures predict prognosis and drug response of HCC patients\u003c/h2\u003e \u003cp\u003eBulk tissue CNA signatures has been reported to be associated with cancer clinical outcomes\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e. To compare the prognostic impact of single-cell CNA signatures with those derived from WGS and SNP array CNA signatures, we analyzed HCC samples with survival data from the PCAWG and TCGA cohorts. We calculated the activity of scSig, WGS_Sig, and SNP_Sig in these samples through signature filtering. Across datasets from different technological platforms, scSig show more robust performance compared with WGS_Sig, and SNP_Sig as reflected in the consistent prognostic outcomes in different datasets (Fig.\u0026nbsp;\u003cspan refid=\"Fig19\" class=\"InternalRef\"\u003e7\u003c/span\u003eA). In two independent datasets, scSig1 and scSig4 were associated with better prognosis, while scSig3 was related to poorer outcomes in the SNP dataset, with a similar trend observed in the WGS dataset (Fig.\u0026nbsp;\u003cspan refid=\"Fig19\" class=\"InternalRef\"\u003e7\u003c/span\u003eB and Figure S13A-B). This finding suggests the superiority of single-cell CNA signatures over traditional bulk tissue CNA signatures in terms of HCC prognosis prediction.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWe analyzed the correlations between drug sensitivity and CNA signatures. Our findings indicate that the sensitivity of HCC patients to the commonly used drug sorafenib is negatively correlated with the scSig7. Additionally, certain drugs demonstrated different responses to particular CNA signatures. Inhibitors of protein arginine methyltransferase (PRMT) have been found to impair tumorigenesis in HCC mouse models\u003csup\u003e\u003cspan additionalcitationids=\"CR50\" citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e. GSK591, a PRMT5 inhibitor, is regulated by MYC in HCC and has been identified as a direct target gene of MYC\u003csup\u003e\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e,\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e. Patients carrying the scSig1 signature exhibited heightened sensitivity to GSK591, whereas they were non-responsive or showed resistance to other drugs (Figure S13C). Importantly, the scSig3 pattern, indicative of normal genomic features, was sensitive to most drugs, suggesting that early-stage tumors without significant CNAs have not yet developed substantial drug resistance. As genomic CNAs increase, tumors gradually acquire drug resistance. These results provide a new perspective for using single-cell CNA signatures in studying tumor drug resistance.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we have developed an analytical framework for the extraction of CNA signatures in single cells. This framework is adaptable across various experimental platforms and sequencing depths, effectively extracting CNA signatures.\u003c/p\u003e \u003cp\u003eCNA signatures have demonstrated their clinical value in a multitude of cancer studies\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. Nevertheless, these existing CNA signature analyzes are all performed using bulk tissue samples, and single-cell CNA signature analysis is needed. We have systematically compared CNA signatures in single cells with those in bulk tissues of HCC. This study demonstrates that the majority of CNA signatures identified at the single-cell level correspond closely with those found in bulk tissue sequencing data of HCC. By employing existing methods to simulate CNA mutational processes and CNA noise, this research evaluated the reliability of different CNA signatures. The findings indicate that single-cell CNA signatures reveal potential CNA mutational processes with higher precision and exhibit greater stability in the presence of simulated noise. Despite the high heterogeneity of HCC, single-cell data accurately reflects the true state of CNA, as it is not affected by the mixing of subclones with different CNA.\u003c/p\u003e \u003cp\u003eThis study is the first attempt to depict the dynamic changes of CNA signatures during HCC progression. We found that CNA signatures with small fragment amplifications, such as scSig7, typically appear in the early stages of HCC development. In contrast, CNA signatures indicative of chromosomal and genomic-level changes, such as scSig1 and scSig2, tend to emerge in the later stages. A reasonable explanation for this pattern is that amplification of small DNA fragments in the early stages disrupts normal cell cycle processes, subsequently leading to large-scale genomic variations, which contribute to the formation of complex large CNA patterns. However, these conclusions still require direct experimental evidence for further validation.\u003c/p\u003e \u003cp\u003eCNA is a complex biological process, and studies focusing on bulk tissues have already demonstrated its potential clinical utility. Nevertheless, the biological significance of certain CNA patterns remains to be elucidated, necessitating further in-depth investigation. In this study, we show that CNA signatures derived from single cells can effectively predict patient prognosis and drug sensitivity. Combining single-cell data with bulk tissue CNA data for integrated analysis could be a promising avenue for future research.\u003c/p\u003e \u003cp\u003eFrom a single-cell perspective, we have identified some potential issues with bulk tissue CNA signatures. However, our study also has certain limitations. Although this study included data from 1,222 single cells, these data were derived from only 10 patients. The limited sample size might have led to the failure to detect all CNA features. Regarding the evolutionary trajectory analysis of CNA signatures, our analysis suggest that certain signatures consistently appear earlier in different samples. However, the complexity of the CNA process makes it challenging to establish an accurate evolutionary timeline and comprehensively describe the detailed evolutionary trajectory of CNA signatures. These analyses provide a direction for future research, and more experimental evidence is required for validation.\u003c/p\u003e\n\u003ch3\u003eResource availability\u003c/h3\u003e\n\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\n \u003ch2\u003eLead contact\u003c/h2\u003e\n \u003cp\u003eFurther information and requests for resources should be directed to and will be fulfilled by the lead contact, Xue-Song Liu ([email protected])\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\n \u003ch2\u003eMaterials availability\u003c/h2\u003e\n \u003cp\u003eThis study did not generate new materials.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\n \u003ch2\u003eData and code availability\u003c/h2\u003e\n \u003cp\u003eOnly publicly available data were used in this study, and data sources and handling of these data are described in the Materials and Methods. All codes required to reproduce the results reported in this manuscript are freely available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/XSLiuLab/single-cell-CNA-signature\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e\n \u003cp\u003eFurther information is available from the corresponding author upon request.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Method details","content":"\u003ch2\u003eData collection and processing\u003c/h2\u003e\u003cp\u003eWe downloaded data from 1222 single-cell shallow whole-genome sequencing (0.4×) of HCC patient samples (totaling 10 patients) from the Genome Sequence Archive (GSA)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e, under accession numbers HRA000094. Data from 160 HCC bulk samples with shallow whole-genome sequencing (0.5×) were also obtained from the International Cancer Genome Consortium (ICGC) database\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e. Moreover, absolute copy number data of whole-genome sequencing from 178 HCC bulk samples in the Pan-Cancer Analysis of Whole Genomes (PCAWG) cohort were acquired through the UCSC Xena platform\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e29\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e. Concurrently, SNP array data from 371 HCC bulk samples were sourced from The Cancer Genome Atlas (TCGA) database\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eFor genome sequencing raw data (fastq) files, fastp was first used to perform quality control on the data, the BWA MEM algorithm was then applied using hg38 as reference genome\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e54\u003c/span\u003e\u003c/sup\u003e, and finally, the samtools tool was used to convert the SAM files into BAM files\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e55\u003c/span\u003e\u003c/sup\u003e. After obtaining the BAM files for all datasets, use the SortSam and MarkDuplicates commands in the Picard toolkit (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://broadinstitute.github.io/picard/\u003c/span\u003e\u003c/span\u003e) to sort BAM files and mark duplications for variant and copy number calling. Raw data processing the codes used can be found at Code availability.\u003c/p\u003e\u003ch2\u003eAbsolute copy number calling\u003c/h2\u003e\u003cp\u003eQDNAseq\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e56\u003c/span\u003e\u003c/sup\u003e and ACE\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e R packages were used to determine the copy number profiles from the low-coverage WGS BAM files. To determine the autosomal copy number profiles, we excluded both chromosome X\u0026amp;Y and mitochondrial DNA and the ploidy was adjusted using median bin segment value, which was the central assumption of ACE. For single-cell sequencing files, we used a ploidy penalty of 0.5 and lower-cellularity penalty of 1.0 to fit the “squaremodel()” function of ACE. For bulk tissue sequencing files, we used a ploidy penalty of 0.5 and lower-cellularity penalty of 0.5 to fit the “squaremodel()” function of ACE as per the author’s recommendation. ABSOLUTE\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e58\u003c/span\u003e\u003c/sup\u003e was used to determine the copy number profiles from the SNP array files of TCGA. The code for copy number calling can be found in Code availability.\u003c/p\u003e\u003ch2\u003eCNA feature classification\u003c/h2\u003e\u003cp\u003eConsidering that shallow sequencing data cannot assess the LOH status of the genome, in order to describe the copy number variation of the liver cancer genome more accurately, we classify the copy number segments from four aspects, (1): copy number, which directly reflects the copy number amplification or deletion, divided into 6 states: 0, 1, 2, 3, 4, 5+, where 0 and 1 indicate that the genome fragment is missing, 2 indicates the normal genome state, and 3 and 4 indicate that the corresponding genome fragment is moderately amplified., 5 + is considered highly amplified. (2): segment length, indicating the length of the copy number segment. The segment length of some small genomic variations such as ecDNA is mostly less than 5Mb\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e59\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e60\u003c/span\u003e\u003c/sup\u003e. Since most of the copy number variations produced by large scale transitions are greater than 10MB\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e61\u003c/span\u003e\u003c/sup\u003e, the copy number segment is length is divided into three types: S (5Mb\u0026lt;), M (5 ~ 10Mb), and L (\u0026gt; 10Mb). (3): segment change, indicating the change in copy number between two segments. Different mutations may imply that the mutations in adjacent segments come from different mechanisms. Divided into two types: AA ( \u0026lt; = 2) and BB (\u0026gt; 2). (4): segment shape, the relationship between the copy number segment and two adjacent segments, divided into three types: HH (left High and right High), LL (left Low and right Low), and OT (other status).\u003c/p\u003e\u003ch2\u003eDeriving signatures\u003c/h2\u003e\u003cp\u003eCopy number signatures were extracted from individual sample matrices using the gold standard tool SigProfiler v1.0.17 (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/AlexandrovLab/SigProfilerExtractor\u003c/span\u003e\u003c/span\u003e) with default parameters. Initially, the non-negative matrix factorization (NMF) algorithm was applied for the decomposition of the patient's copy number feature matrix. Signature search intervals ranging from 3 to 12 were set, and NMF was run 200 times for each signature number with different random seeds to ensure result stability. For each NMF iteration, initialization was performed using random numbers, and iterations ranged from 10,000 to 1,000,000 until stable results were obtained. Two key parameters determined during this process were the stability of the signatures, as measured by silhouette analysis, and the average Frobenius reconstruction error. Under these criteria, seven signatures were identified as optimal at the single-cell level and were selected for further analysis.\u003c/p\u003e\u003ch2\u003eCopy number variation simulation\u003c/h2\u003e\u003cp\u003eTo simulate realistic chromosomal variation events, this study utilized the approach proposed by Drews et al\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. We simulated WGD events by setting a tetraploid background prior to the insertion of other CNAs for early WGD, or by multiplying all copy number states by a factor following the placement of all CNAs for late WGD. Considering that most WGD events are early occurrences, the ratio of early to late WGD samples was set at 4:1\u003csup\u003e62\u003c/sup\u003e. The simulation of LST was based on their real breakpoint characteristics, employing a Poisson distribution for this simulation. Random copy number variations were simulated using the RSVSim\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e tool, with hg38 as the reference genome. The simulations aimed to cover 0–15% of the genome. The length of each amplified and deleted segment was randomly set between 1kb and 5MB, distributed randomly across the entire genome, with all other parameters set to their default values. Calculating the degree of similarity between simulated samples and scSig using cosine similarity, a method widely used for assessing similarities among copy number profiles\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003ch2\u003ePhylogenic trees\u003c/h2\u003e\u003cp\u003eWe utilized the phangorn package in R to calculate the Maximum Parsimony Tree from the scSig matrix using the Parsimony Ratchet algorithm\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e63\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e64\u003c/span\u003e\u003c/sup\u003e. To ensure the accuracy of our analysis, samples that did not exhibit any CNAs were excluded. The branch lengths between cells were estimated using the ACCTRAN algorithm. The \"ggtree\" package in R was employed to visualize the phylogenetic trees\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e65\u003c/span\u003e\u003c/sup\u003e, where the predominant CNA signature were marked with distinct colors for easy visual differentiation.\u003c/p\u003e\u003ch2\u003eCNA signature evolution\u003c/h2\u003e\u003cp\u003eTo determine the temporal sequence of emergence for tumor cell subclones, we estimated the likely order by analyzing shared CNA breakpoints. Shared breakpoints refer to genomic CNAs that occur at the same positions across different subclonal cells within the same sample. Each subclone’s CNA may build upon the segments from prior CNA events, but these subclones are likely to retain the breakpoints from earlier CNAs.\u003c/p\u003e\u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u0026nbsp;\u003cimg src=\"data:image/png;base64,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\"\u003e\u003c/span\u003e\u003c/p\u003e\u003cp\u003ewhere S\u003csub\u003ej\u003c/sub\u003e is the frequency of shared break point of cell j in the sample, M is the number of cells in the sample, L\u003csub\u003ej\u003c/sub\u003e is the number of shared breakpoints in cell j, and I\u003csub\u003eij\u003c/sub\u003e represents the existence of the shared breakpoint I in cell j. I\u003csub\u003eij\u003c/sub\u003e is a binary variable. When the breakpoint i exists in the cell j, I\u003csub\u003eij\u003c/sub\u003e takes a value of 1; when the breakpoint i does not exist in the cell j, I\u003csub\u003eij\u003c/sub\u003e takes a value of 0.\u003c/p\u003e\u003ch2\u003eSurvival analysis\u003c/h2\u003e\u003cp\u003eIn this study, we gathered survival data for samples from the TCGA (SNP array sequencing) and PCAWG (WGS sequencing) projects, encompassing 352 samples from TCGA and 139 samples from PCAWG. R package sigminer\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e was utilized to fit each patient's copy number feature matrix to scSig, SNP_Sig, and WGS_Sig, respectively. The association between CNA feature activity and overall survival was determined using univariate Cox proportional hazards models. For each Cox model, Hazard ratio were calculated and reported to indicate the direction and significance of the survival relationship. The Cox model is given by:\u003c/p\u003e\u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u0026nbsp;\u003cimg src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAKYAAAAuCAYAAACia2LCAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsMAAA7DAcdvqGQAAAgZSURBVHhe7Z05rA5fFMDv+0dpKSiISCxBFLYIFYmtsEXHExGEoFDYngQVncJaSXgSFJYobYUlCoUQSygEQSGWBIWlf//5nTfn5b4x33x3Vvcl95dMvjd3tnPPPfecc8+M6OiJMIGAZ/wX/wYCXhEMM+AlwTADXhIMM+AlwTADXhIMM+AlwTADXhIMM+AlwTADXhIMM+AlwTADXtK4Yb59+9acPXvW/PnzJ275tyDH8ePHzZcvX+KWgY+Ljuvq982bN82cOXNMR0eH/HJ/tgkTJkgbv+xfv35d9tk2bNgQX23BRxxN8ebNm56urq6e379/xy1+gDzIhXwDnTw6rrrf165d65k9e7bcj3svWrSo58iRI3Ls8+fPcuzp06d9+1nPbswwEaSzs1N+fQSFLVmyxFv5XCii4yr7vX79ejFOSJMFI2Wjrbu7O25Np7FQvm/fPhPNGDNq1Ki4xS9mzpxpFi9eLHIOVIrouKp+P3v2zLx69coMGzZMQvimTZskVbBlmTx5srly5YqE+82bN8et6TgbJjmJ5g90JE+OiNAPHz40CxcujFv8BPmQE3mbwM692I4ePRofyU8ZHVfR73v37pkpU6aYefPmmUePHomNXLx4MT7ay6xZs8Rwly1bFre0xtkwt2/fbiJXbR4/fhy3uHPixAkzZswYM3HixLilWhhQHVzdNMkGFD5ixIh+x9MmF/IhJ/I2Ad7k3bt3JgpvZvjw4aUmbhkdl+03erx165ZZtWpV3NLLixcv4r96z4nCtxk0aJD5+vVr3JpBHNKdIB+JFChJresChnxi/PjxfUlwXahsdEnzHBs9vmXLFvm7FciJvFXkXC7oIqHMM6vQcZl+o08WNnqt9knHgf3I6OWX57jIWXuO+eTJE/P+/XvJL7LA61FCKAq50v79++Xvq1evyq/CbD18+LC5cOGCOX36tJzbCuREXuRugmiwzIcPH8zcuXML59+uOoZWei7Tb8L40KFDRc9sPGP69OlmxYoVUroirySnHDx4sEQFQj3nZRIbqBNFPCYrNa7J8lLMrCq8VJp8/B48eDDz+TZ6D+RuAvrOMKR5eVdcdAxZei7ab/SLvs+cOSNek3scO3asJ0qjpJ2+6TPZ+Ju2dvKW8piXL1/uy91YFCWTZ2bFp0+f4r10yDs2btxYymMo5EozZswwz58/l5mqs3flypWZXjIN5G47qysA7x4Nkhk9enTckg8XHYOrnvP2Gz3/+vXLLF++XDzh9+/fza5du8zIkSPNnTt3cHySR/NMzalp47zMMRHzdERnFTPDLo5qe9KT6gxJm6V6DSLYW5k8CZjxep88nlLJkrlq1NvYertx44bolz4k9ZlGO3ld9Zx1H8aZWqdei3yqV9ecMS+FDdMWXhXMMdsQ2ilNr2t1vAgaGpOyuKIyF70+D6pPHVjCIWEQvTDBXPTSTsfgoudW/d69e7e0Xbp0SfY5xnnIxzW2kVZJoVBOojtkyJB4z0hSmxaKKAv8/Pkz3vsbwgBht4owrlArixQn9TLCSZWklaWyNs7PgkXDjx8/ZOFBqEVewiAQUseNG9dPz2m00zEU1fOePXvM+fPnze3bt82aNWviViPPmzZtmtyL8J03TXKh0lU5SnbJdxQdmGT9qwysuqmV1bGyZqCiyey8cX4W1PmYRPfv35fBZhULakhMdiZ9WYromZV75L3NuXPn+gwPubZt2yYyr127VtrqotZyER4Az5UGCTZF2ShMFE78kxw6dEgWOlu3bpX9169fy6+PUPx/8OCBTKBv3771M+I8hpSlYyiiZ66h2I4BEoFY1KLTSZMmyZsd5KsqwrWi9jpmK6K8R+p3rKKreCPEwKJEZje1MgaCAcmzwrSpIxWw0dojg0+NVVFDUqMoS14902+uwWMj39SpU82BAwfkdSPtpCdVePF21GqY5EfkSYSp5GsoHZilS5eW7ihGuWDBgr5QmCwbFSEtv6syx1RvfvLkyX7eRw1J80F9ncqW9i47S8eQV8/ci/IPHjta4EhZh4lC7tuEQSq1GiYdaRU+tH6Hd8NLOL0/TcB1e/fuNevWreszStDn5s15QRcTafldVTmmesVopWzmz58ft/aihkQYf/nypXylw+KDjfxO3/8rWToGVz2n9Tutpsnk6Ozs/Ku9anIZJkpjsJnRtoIQUg0gmdeh4KSBaH6FV8Mr8HmUbQTqmSjap3k8nsc1Y8eOlXCbtipk1Qg7duzI5TWR0zW/K0rW4gZDIowTUvkKnX7QP6IA1RDGIEmajqGdnm3sfjNZmDR3796VsUDfbFQOTp06JXK1uk9lRDO7LVF4kToYp9sbtSxqcMn2SLF99TKtj9lFWNq4XzST5aMK7m9DYZz7RMn2X6/I0p5nn6PPS57D85LPSYP72/LXQavXkMhHX9ALNU1ksfum+k6SpmNop2ebZL8pqmuhn40CO8X/pnAyzLKgUFfDsEFJGGmTIKttDP8SjMU2tlaGCUV1rPjUb2hkVb5z507z8ePHXCGV0EFNUks/TUDo4yts5B1oFNGx4mO/GzFMciQMjPqXC3wpz6dSLB7qrpfZ8MU1ZZG0nPVfwBsh/dhW8/hWn7bl1bGNb/0WYs9ZO4SY1atX1/7+uSikDb79Yzlk0XfRbOgvK1QX0bGP/YbGDBN8VQIDytdSPk4aZGLxwuYiXx4d+9zvRg0TUEZ3d3fmzG8SlafVv28eiLjo2Pd+h/+1IuAljSx+AoG8BMMMeEkwzICXBMMMeEkwzICXBMMMeEkwzICHGPM/3oPTxsKn9qQAAAAASUVORK5CYII=\"\u003e\u003c/span\u003e\u003c/p\u003e\u003cp\u003ewhere t is the total survival time, h(t, X) is the hazard function given the presence of a prognostic variable X, and h0(t) represents the baseline hazard. Cox proportional hazards analysis was conducted using the R package ezcox\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e66\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003ch2\u003eDrug sensitivity analysis\u003c/h2\u003e\u003cp\u003eGenomic information and drug sensitivity data for all HCC cell lines were downloaded from the GDSC database\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e67\u003c/span\u003e\u003c/sup\u003e. Following the previously introduced method, copy number variation data were obtained, and copy number feature matrices were calculated. Subsequently, using the R package sigminer, the copy number feature matrices of each cell line were fitted to scSig. Spearman correlation analysis was utilized to assess the relationship between the exposure values of scSig for each sample and IC50. The significance P-values obtained were corrected using the FDR method. Finally, the R package ComplexHeatmap\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e68\u003c/span\u003e\u003c/sup\u003e was employed to draw correlation heatmap.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank ShanghaiTech University High Performance Computing Public Service Platform for computing services. We thank multi-omics facility, molecular and cell biology core facility of ShanghaiTech University for technical help. This work is supported by cross disciplinary Research Fund of Shanghai Ninth People\u0026apos;s Hospital, Shanghai JiaoTong University School of Medicine (JYJC202227). Shanghai Science and Technology Commission (21ZR1442400), National Natural Science Foundation of China (82373149), and startup funding from ShanghaiTech University.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eC.W. performed all the CNA signature analysis and drafted the manuscript under the supervision of X.-S. L. \u0026nbsp;Z.T., W.C., N.W., J.W., J.S., T.W., D.Q., K.D., X.Z. T.L., L.Z., W.W., X.L., X.L., participated in cancer genome analysis and discussion. \u0026nbsp;X.X. provided critical materials for this study and participated in critical project discussion and supervision. X.-S. L. conceived this study, supervised this study and wrote the manuscript with C.W.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eDeclaration of interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe declare no competing interest.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eHanahan, D. (2022). Hallmarks of Cancer: New Dimensions. Cancer discovery \u003cem\u003e12\u003c/em\u003e, 31-46. 10.1158/2159-8290.Cd-21-1059.\u003c/li\u003e\n\u003cli\u003eAlexandrov, L.B., Nik-Zainal, S., Wedge, D.C., Aparicio, S.A.J.R., Behjati, S., Biankin, A.V., Bignell, G.R., Bolli, N., Borg, A., B\u0026oslash;rresen-Dale, A.-L., et al. (2013). Signatures of mutational processes in human cancer. Nature \u003cem\u003e500\u003c/em\u003e, 415-421. 10.1038/nature12477.\u003c/li\u003e\n\u003cli\u003ePoulsgaard, G.A., S\u0026oslash;rensen, S.G., Juul, R.I., Nielsen, M.M., and Pedersen, J.S. (2023). Sequence dependencies and mutation rates of localized mutational processes in cancer. Genome Medicine \u003cem\u003e15\u003c/em\u003e, 63. 10.1186/s13073-023-01217-z.\u003c/li\u003e\n\u003cli\u003eYang, G., Shahatiaili, A., Bai, S., Wang, L., Jin, D., Cao, M., Su, P., Liu, Q., Tao, K., and Long, Q.J.T.J.o.P. (2024). Mutational signature and prognosis in adenocarcinoma of the bladder. \u003cem\u003e262\u003c/em\u003e, 334-346.\u003c/li\u003e\n\u003cli\u003eAlexandrov, L.B., Kim, J., Haradhvala, N.J., Huang, M.N., Tian Ng, A.W., Wu, Y., Boot, A., Covington, K.R., Gordenin, D.A., Bergstrom, E.N., et al. (2020). The repertoire of mutational signatures in human cancer. Nature \u003cem\u003e578\u003c/em\u003e, 94-101. 10.1038/s41586-020-1943-3.\u003c/li\u003e\n\u003cli\u003eSteele, C.D., Abbasi, A., Islam, S.A., Bowes, A.L., Khandekar, A., Haase, K., Hames-Fathi, S., Ajayi, D., Verfaillie, A., and Dhami, P.J.N. (2022). Signatures of copy number alterations in human cancer. \u003cem\u003e606\u003c/em\u003e, 984-991.\u003c/li\u003e\n\u003cli\u003eDrews, R.M., Hernando, B., Tarabichi, M., Haase, K., Lesluyes, T., Smith, P.S., Morrill Gavarr\u0026oacute;, L., Couturier, D.-L., Liu, L., and Schneider, M.J.N. (2022). A pan-cancer compendium of chromosomal instability. \u003cem\u003e606\u003c/em\u003e, 976-983.\u003c/li\u003e\n\u003cli\u003eTao, Z., Wang, S., Wu, C., Wu, T., Zhao, X., Ning, W., Wang, G., Wang, J., Chen, J., and Diao, K.J.B.i.B. (2023). The repertoire of copy number alteration signatures in human cancer. \u003cem\u003e24\u003c/em\u003e, bbad053.\u003c/li\u003e\n\u003cli\u003eWang, S., Tao, Z., Wu, T., and Liu, X.-S.J.B. (2021). Sigflow: an automated and comprehensive pipeline for cancer genome mutational signature analysis. \u003cem\u003e37\u003c/em\u003e, 1590-1592.\u003c/li\u003e\n\u003cli\u003eWang, S., Jia, M., He, Z., and Liu, X.-S.J.O. (2018). APOBEC3B and APOBEC mutational signature as potential predictive markers for immunotherapy response in non-small cell lung cancer. \u003cem\u003e37\u003c/em\u003e, 3924-3936.\u003c/li\u003e\n\u003cli\u003eMacintyre, G., Goranova, T.E., De Silva, D., Ennis, D., Piskorz, A.M., Eldridge, M., Sie, D., Lewsley, L.-A., Hanif, A., Wilson, C., et al. (2018). Copy number signatures and mutational processes in ovarian carcinoma. Nature Genetics \u003cem\u003e50\u003c/em\u003e, 1262-1270. 10.1038/s41588-018-0179-8.\u003c/li\u003e\n\u003cli\u003eDoran, C.G., and Pennington, S.R.J.B.i.M. (2022). Copy number alteration signatures as biomarkers in cancer: a review. \u003cem\u003e16\u003c/em\u003e, 371-386.\u003c/li\u003e\n\u003cli\u003eEssers, P.B., van der Heijden, M., Vossen, D., de Roest, R.H., Leemans, C.R., Brakenhoff, R.H., van den Brekel, M.W., Bartelink, H., Verheij, M., and Vens, C.J.I.j.o.c. (2020). Ovarian cancer‐derived copy number alterations signatures are prognostic in chemoradiotherapy‐treated head and neck squamous cell carcinoma. \u003cem\u003e147\u003c/em\u003e, 1732-1739.\u003c/li\u003e\n\u003cli\u003eYao, H., Li, H., Wang, J., Wu, T., Ning, W., Diao, K., Wu, C., Wang, G., Tao, Z., and Zhao, X.J.C.B. (2023). Copy number alteration features in pan-cancer homologous recombination deficiency prediction and biology. \u003cem\u003e6\u003c/em\u003e, 527.\u003c/li\u003e\n\u003cli\u003eWang, S., Li, H., Song, M., Tao, Z., Wu, T., He, Z., Zhao, X., Wu, K., and Liu, X.-S.J.P.G. (2021). Copy number signature analysis tool and its application in prostate cancer reveals distinct mutational processes and clinical outcomes. \u003cem\u003e17\u003c/em\u003e, e1009557.\u003c/li\u003e\n\u003cli\u003eNguyen, P.H.D., Ma, S., Phua, C.Z.J., Kaya, N.A., Lai, H.L.H., Lim, C.J., Lim, J.Q., Wasser, M., Lai, L., Tam, W.L., et al. (2021). Intratumoural immune heterogeneity as a hallmark of tumour evolution and progression in hepatocellular carcinoma. Nature Communications \u003cem\u003e12\u003c/em\u003e, 227. 10.1038/s41467-020-20171-7.\u003c/li\u003e\n\u003cli\u003eLi, L., and Wang, H.J.C.l. (2016). Heterogeneity of liver cancer and personalized therapy. \u003cem\u003e379\u003c/em\u003e, 191-197.\u003c/li\u003e\n\u003cli\u003eHiley, C., de Bruin, E.C., McGranahan, N., and Swanton, C.J.G.b. (2014). Deciphering intratumor heterogeneity and temporal acquisition of driver events to refine precision medicine. \u003cem\u003e15\u003c/em\u003e, 1-10.\u003c/li\u003e\n\u003cli\u003eLei, H., Lyu, B., Gertz, E.M., Sch\u0026auml;ffer, A.A., Shi, X., Wu, K., Li, G., Xu, L., Hou, Y., and Dean, M.J.J.o.C.B. (2020). Tumor copy number deconvolution integrating bulk and single-cell sequencing data. \u003cem\u003e27\u003c/em\u003e, 565-598.\u003c/li\u003e\n\u003cli\u003eMalikic, S., Mehrabadi, F.R., Ciccolella, S., Rahman, M.K., Ricketts, C., Haghshenas, E., Seidman, D., Hach, F., Hajirasouliha, I., and Sahinalp, S.C. (2019). PhISCS: a combinatorial approach for subperfect tumor phylogeny reconstruction via integrative use of single-cell and bulk sequencing data. Genome research \u003cem\u003e29\u003c/em\u003e, 1860-1877. 10.1101/gr.234435.118.\u003c/li\u003e\n\u003cli\u003eGraham, T.A., and Sottoriva, A.J.T.J.o.p. (2017). Measuring cancer evolution from the genome. \u003cem\u003e241\u003c/em\u003e, 183-191.\u003c/li\u003e\n\u003cli\u003eTurajlic, S., Sottoriva, A., Graham, T., and Swanton, C.J.N.R.G. (2019). Resolving genetic heterogeneity in cancer. \u003cem\u003e20\u003c/em\u003e, 404-416.\u003c/li\u003e\n\u003cli\u003eGao, R., Davis, A., McDonald, T.O., Sei, E., Shi, X., Wang, Y., Tsai, P.-C., Casasent, A., Waters, J., and Zhang, H.J.N.g. (2016). Punctuated copy number evolution and clonal stasis in triple-negative breast cancer. \u003cem\u003e48\u003c/em\u003e, 1119-1130.\u003c/li\u003e\n\u003cli\u003eGuo, L., Yi, X., Chen, L., Zhang, T., Guo, H., Chen, Z., Cheng, J., Cao, Q., Liu, H., Hou, C., et al. (2022). Single-Cell DNA Sequencing Reveals Punctuated and Gradual Clonal Evolution in Hepatocellular Carcinoma. Gastroenterology \u003cem\u003e162\u003c/em\u003e, 238-252. 10.1053/j.gastro.2021.08.052.\u003c/li\u003e\n\u003cli\u003eGerstung, M., Jolly, C., Leshchiner, I., Dentro, S.C., Gonzalez, S., Rosebrock, D., Mitchell, T.J., Rubanova, Y., Anur, P., and Yu, K.J.N. (2020). The evolutionary history of 2,658 cancers. \u003cem\u003e578\u003c/em\u003e, 122-128.\u003c/li\u003e\n\u003cli\u003eMcPherson, A., Roth, A., Laks, E., Masud, T., Bashashati, A., Zhang, A.W., Ha, G., Biele, J., Yap, D., and Wan, A.J.N.g. (2016). Divergent modes of clonal spread and intraperitoneal mixing in high-grade serous ovarian cancer. \u003cem\u003e48\u003c/em\u003e, 758-767.\u003c/li\u003e\n\u003cli\u003eDatabase Resources of the National Genomics Data Center, China National Center for Bioinformation in 2023. (2023). Nucleic acids research \u003cem\u003e51\u003c/em\u003e, D18-d28. 10.1093/nar/gkac1073.\u003c/li\u003e\n\u003cli\u003eZhang, J., Bajari, R., Andric, D., Gerthoffert, F., Lepsa, A., Nahal-Bose, H., Stein, L.D., and Ferretti, V.J.N.b. (2019). The international cancer genome consortium data portal. \u003cem\u003e37\u003c/em\u003e, 367-369.\u003c/li\u003e\n\u003cli\u003ePan-cancer analysis of whole genomes %J Nature. (2020). \u003cem\u003e578\u003c/em\u003e, 82-93.\u003c/li\u003e\n\u003cli\u003eGoldman, M.J., Craft, B., Hastie, M., Repečka, K., McDade, F., Kamath, A., Banerjee, A., Luo, Y., Rogers, D., Brooks, A.N., et al. (2020). Visualizing and interpreting cancer genomics data via the Xena platform. Nature Biotechnology \u003cem\u003e38\u003c/em\u003e, 675-678. 10.1038/s41587-020-0546-8.\u003c/li\u003e\n\u003cli\u003eZack, T.I., Schumacher, S.E., Carter, S.L., Cherniack, A.D., Saksena, G., Tabak, B., Lawrence, M.S., Zhang, C.-Z., Wala, J., and Mermel, C.H.J.N.g. (2013). Pan-cancer patterns of somatic copy number alteration. \u003cem\u003e45\u003c/em\u003e, 1134-1140.\u003c/li\u003e\n\u003cli\u003eKorbel, J.O., and Campbell, P.J.J.C. (2013). Criteria for inference of chromothripsis in cancer genomes. \u003cem\u003e152\u003c/em\u003e, 1226-1236.\u003c/li\u003e\n\u003cli\u003ePopova, T., Mani\u0026eacute;, E., Rieunier, G., Caux-Moncoutier, V., Tirapo, C., Dubois, T., Delattre, O., Sigal-Zafrani, B., Bollet, M., and Longy, M.J.C.r. (2012). Ploidy and large-scale genomic instability consistently identify basal-like breast carcinomas with BRCA1/2 inactivation. \u003cem\u003e72\u003c/em\u003e, 5454-5462.\u003c/li\u003e\n\u003cli\u003eAl-Rawi, D.H., Bakhoum, S.F.J.C.O.i.G., and Development (2022). Chromosomal instability as a source of genomic plasticity. \u003cem\u003e74\u003c/em\u003e, 101913.\u003c/li\u003e\n\u003cli\u003eMenghi, F., Inaki, K., Woo, X., Kumar, P.A., Grzeda, K.R., Malhotra, A., Yadav, V., Kim, H., Marquez, E.J., and Ucar, D.J.P.o.t.N.A.o.S. (2016). The tandem duplicator phenotype as a distinct genomic configuration in cancer. \u003cem\u003e113\u003c/em\u003e, E2373-E2382.\u003c/li\u003e\n\u003cli\u003eAlexandrov, L.B., Nik-Zainal, S., Wedge, D.C., Aparicio, S.A., Behjati, S., Biankin, A.V., Bignell, G.R., Bolli, N., Borg, A., and B\u0026oslash;rresen-Dale, A.-L.J.N. (2013). Signatures of mutational processes in human cancer. \u003cem\u003e500\u003c/em\u003e, 415-421.\u003c/li\u003e\n\u003cli\u003eAlexandrov, L.B., Kim, J., Haradhvala, N.J., Huang, M.N., Tian Ng, A.W., Wu, Y., Boot, A., Covington, K.R., Gordenin, D.A., and Bergstrom, E.N.J.N. (2020). The repertoire of mutational signatures in human cancer. \u003cem\u003e578\u003c/em\u003e, 94-101.\u003c/li\u003e\n\u003cli\u003eKim, H., Nguyen, N.-P., Turner, K., Wu, S., Gujar, A.D., Luebeck, J., Liu, J., Deshpande, V., Rajkumar, U., and Namburi, S.J.N.g. (2020). Extrachromosomal DNA is associated with oncogene amplification and poor outcome across multiple cancers. \u003cem\u003e52\u003c/em\u003e, 891-897.\u003c/li\u003e\n\u003cli\u003eYi, E., Chamorro Gonz\u0026aacute;lez, R., Henssen, A.G., and Verhaak, R.G.W. (2022). Extrachromosomal DNA amplifications in cancer. Nature Reviews Genetics \u003cem\u003e23\u003c/em\u003e, 760-771. 10.1038/s41576-022-00521-5.\u003c/li\u003e\n\u003cli\u003eZhao, X.-K., Xing, P., Song, X., Zhao, M., Zhao, L., Dang, Y., Lei, L.-L., Xu, R.-H., Han, W.-L., and Wang, P.-P.J.N.c. (2021). Focal amplifications are associated with chromothripsis events and diverse prognoses in gastric cardia adenocarcinoma. \u003cem\u003e12\u003c/em\u003e, 6489.\u003c/li\u003e\n\u003cli\u003eKoche, R.P., Rodriguez-Fos, E., Helmsauer, K., Burkert, M., MacArthur, I.C., Maag, J., Chamorro, R., Munoz-Perez, N., Puiggros, M., and Dorado Garcia, H.J.N.g. (2020). Extrachromosomal circular DNA drives oncogenic genome remodeling in neuroblastoma. \u003cem\u003e52\u003c/em\u003e, 29-34.\u003c/li\u003e\n\u003cli\u003eDentro, S.C., Leshchiner, I., Haase, K., Tarabichi, M., Wintersinger, J., Deshwar, A.G., Yu, K., Rubanova, Y., Macintyre, G., and Demeulemeester, J.J.C. (2021). Characterizing genetic intra-tumor heterogeneity across 2,658 human cancer genomes. \u003cem\u003e184\u003c/em\u003e, 2239-2254. e2239.\u003c/li\u003e\n\u003cli\u003eDavoli, T., de Lange, T.J.A.r.o.c., and biology, d. (2011). The causes and consequences of polyploidy in normal development and cancer. \u003cem\u003e27\u003c/em\u003e, 585-610.\u003c/li\u003e\n\u003cli\u003eBartenhagen, C., and Dugas, M.J.B. (2013). RSVSim: an R/Bioconductor package for the simulation of structural variations. \u003cem\u003e29\u003c/em\u003e, 1679-1681.\u003c/li\u003e\n\u003cli\u003eGiguere, C., Dubey, H.V., Sarsani, V.K., Saddiki, H., He, S., and Flaherty, P.J.B.b. (2020). SCSIM: Jointly simulating correlated single-cell and bulk next-generation DNA sequencing data. \u003cem\u003e21\u003c/em\u003e, 1-10.\u003c/li\u003e\n\u003cli\u003eKaufmann, T.L., Petkovic, M., Watkins, T.B., Colliver, E.C., Laskina, S., Thapa, N., Minussi, D.C., Navin, N., Swanton, C., and Van Loo, P. (2022). MEDICC2: whole-genome doubling aware copy-number phylogenies for cancer evolution. Genome biology \u003cem\u003e23\u003c/em\u003e, 241.\u003c/li\u003e\n\u003cli\u003eComaills, V., and Castellano-Pozo, M.J.B. (2023). Chromosomal instability in genome evolution: from cancer to macroevolution. \u003cem\u003e12\u003c/em\u003e, 671.\u003c/li\u003e\n\u003cli\u003eMaclachlan, K.H., Rustad, E.H., Derkach, A., Zheng-Lin, B., Yellapantula, V., Diamond, B., Hultcrantz, M., Ziccheddu, B., Boyle, E.M., and Blaney, P.J.N.c. (2021). Copy number signatures predict chromothripsis and clinical outcomes in newly diagnosed multiple myeloma. \u003cem\u003e12\u003c/em\u003e, 5172.\u003c/li\u003e\n\u003cli\u003eLei, Y., Han, P., and Tian, D. (2021). Protein arginine methyltransferases and hepatocellular carcinoma: A review. Translational Oncology \u003cem\u003e14\u003c/em\u003e, 101194.\u003c/li\u003e\n\u003cli\u003eGou, Q., He, S., and Zhou, Z. (2017). Protein arginine N-methyltransferase 1 promotes the proliferation and metastasis of hepatocellular carcinoma cells. Tumor Biology \u003cem\u003e39\u003c/em\u003e, 1010428317691419.\u003c/li\u003e\n\u003cli\u003eJeon, J.-Y., Lee, J.S., Park, E.-R., Shen, Y.N., Kim, M.-Y., Shin, H.-J., Joo, H.-Y., Cho, E.-H., Moon, S.M., and Shin, U.S. (2018). Protein arginine methyltransferase 5 is implicated in the aggressiveness of human hepatocellular carcinoma and controls the invasive activity of cancer cells. Oncology reports \u003cem\u003e40\u003c/em\u003e, 536-544.\u003c/li\u003e\n\u003cli\u003eLuo, Y., Gao, Y., Liu, W., Yang, Y., Jiang, J., Wang, Y., Tang, W., Yang, S., Sun, L., and Cai, J. (2021). Myelocytomatosis‐protein arginine N‐methyltransferase 5 Axis defines the tumorigenesis and immune response in hepatocellular carcinoma. Hepatology \u003cem\u003e74\u003c/em\u003e, 1932-1951.\u003c/li\u003e\n\u003cli\u003eDang, C.V. (2012). MYC on the path to cancer. Cell \u003cem\u003e149\u003c/em\u003e, 22-35.\u003c/li\u003e\n\u003cli\u003eLi, H.J.a.p.a. (2013). Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM.\u003c/li\u003e\n\u003cli\u003eLi, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G., Abecasis, G., Durbin, R., and bioinformatics, G.P.D.P.S.J. (2009). The sequence alignment/map format and SAMtools. \u003cem\u003e25\u003c/em\u003e, 2078-2079.\u003c/li\u003e\n\u003cli\u003eScheinin, I., Sie, D., Bengtsson, H., Van De Wiel, M.A., Olshen, A.B., Van Thuijl, H.F., Van Essen, H.F., Eijk, P.P., Rustenburg, F., and Meijer, G.A.J.G.r. (2014). DNA copy number analysis of fresh and formalin-fixed specimens by shallow whole-genome sequencing with identification and exclusion of problematic regions in the genome assembly. \u003cem\u003e24\u003c/em\u003e, 2022-2032.\u003c/li\u003e\n\u003cli\u003ePoell, J.B., Mendeville, M., Sie, D., Brink, A., Brakenhoff, R.H., and Ylstra, B.J.B. (2019). ACE: absolute copy number estimation from low-coverage whole-genome sequencing data. \u003cem\u003e35\u003c/em\u003e, 2847-2849.\u003c/li\u003e\n\u003cli\u003eCarter, S.L., Cibulskis, K., Helman, E., McKenna, A., Shen, H., Zack, T., Laird, P.W., Onofrio, R.C., Winckler, W., and Weir, B.A.J.N.b. (2012). Absolute quantification of somatic DNA alterations in human cancer. \u003cem\u003e30\u003c/em\u003e, 413-421.\u003c/li\u003e\n\u003cli\u003eWu, S., Turner, K.M., Nguyen, N., Raviram, R., Erb, M., Santini, J., Luebeck, J., Rajkumar, U., Diao, Y., and Li, B.J.N. (2019). Circular ecDNA promotes accessible chromatin and high oncogene expression. \u003cem\u003e575\u003c/em\u003e, 699-703.\u003c/li\u003e\n\u003cli\u003eWu, T., Wu, C., Zhao, X., Wang, G., Ning, W., Tao, Z., Chen, F., and Liu, X.-S.J.S.r. (2022). Extrachromosomal DNA formation enables tumor immune escape potentially through regulating antigen presentation gene expression. \u003cem\u003e12\u003c/em\u003e, 3590.\u003c/li\u003e\n\u003cli\u003eSchonhoft, J.D., Zhao, J.L., Jendrisak, A., Carbone, E.A., Barnett, E.S., Hullings, M.A., Gill, A., Sutton, R., Lee, J., and Dago, A.E.J.C.r. (2020). Morphology-predicted large-scale transition number in circulating tumor cells identifies a chromosomal instability biomarker associated with poor outcome in castration-resistant prostate cancer. \u003cem\u003e80\u003c/em\u003e, 4892-4903.\u003c/li\u003e\n\u003cli\u003eBielski, C.M., Zehir, A., Penson, A.V., Donoghue, M.T., Chatila, W., Armenia, J., Chang, M.T., Schram, A.M., Jonsson, P., and Bandlamudi, C.J.N.g. (2018). Genome doubling shapes the evolution and prognosis of advanced cancers. \u003cem\u003e50\u003c/em\u003e, 1189-1195.\u003c/li\u003e\n\u003cli\u003eParadis, E., and Schliep, K.J.B. (2019). ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R. \u003cem\u003e35\u003c/em\u003e, 526-528.\u003c/li\u003e\n\u003cli\u003eSchliep, K.P.J.B. (2011). phangorn: phylogenetic analysis in R. \u003cem\u003e27\u003c/em\u003e, 592-593.\u003c/li\u003e\n\u003cli\u003eYu, G., Lam, T.T.-Y., Zhu, H., Guan, Y.J.M.b., and evolution (2018). Two methods for mapping and visualizing associated data on phylogeny using ggtree. \u003cem\u003e35\u003c/em\u003e, 3041-3043.\u003c/li\u003e\n\u003cli\u003eWang, S., Liu, X.-S., Li, J., and Zhao, Q.J.a.p.a. (2021). Ezcox: an R/CRAN package for cox model batch processing and visualization.\u003c/li\u003e\n\u003cli\u003eYang, W., Soares, J., Greninger, P., Edelman, E.J., Lightfoot, H., Forbes, S., Bindal, N., Beare, D., Smith, J.A., and Thompson, I.R.J.N.a.r. (2012). Genomics of Drug Sensitivity in Cancer (GDSC): a resource for therapeutic biomarker discovery in cancer cells. \u003cem\u003e41\u003c/em\u003e, D955-D961.\u003c/li\u003e\n\u003cli\u003eGu, Z., and H\u0026uuml;bschmann, D.J.B. (2022). Make interactive complex heatmaps in R. \u003cem\u003e38\u003c/em\u003e, 1460-1462.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"scDNA-seq, copy number signature, HCC evolutionary, Drug sensitivity","lastPublishedDoi":"10.21203/rs.3.rs-6304422/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6304422/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCopy number alteration (CNA) is a major type of cancer genome alteration that drives cancer progression. CNA signature analysis can reveal underlying etiology and provide biomarkers for cancer treatment, and existing CNA signature analyzes are all performed using bulk tissue samples. However CNA usually affect large proportion of genome, and the CNA profile of bulk sample does not reflect the actual CNA profiles of the individual cancer cells of the sample, especially in tumors with high heterogeneity, such as hepatocellular carcinoma (HCC). Furthermore, the evolutionary trajectory of CNA mutational processes still remain elusive. Here we build a method to comprehensively analyze the CNA signatures of HCC from single-cell and bulk sample perspective, revealing patterns and potential noise signals from the usually performed bulk tissue CNA signature analysis. Single-cell signature analysis delineated the evolutionary trajectory of HCC CNA signatures, and different CNA signatures consistently emerge in different HCC evolution stages. Single-cell CNA signatures show robust performance in patient prognosis and drug sensitivity prediction. This work not only reveals specific considerations in analyzing CNA signature derived from bulk tissue but also depicts CNA evolution process and provides potential biomarkers for the prognosis and treatment of HCC patients.\u003c/p\u003e","manuscriptTitle":"Single-cell copy number alteration signature analysis reveals masked patterns and potential biomarkers for cancer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-24 10:08:11","doi":"10.21203/rs.3.rs-6304422/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"communications-biology","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"commsbio","sideBox":"Learn more about [Communications Biology](http://www.nature.com/commsbio/)","snPcode":"","submissionUrl":"","title":"Communications Biology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Communications Series","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"221044a2-e088-49db-9b9f-8693a6cbe841","owner":[],"postedDate":"April 24th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":46630002,"name":"Biological sciences/Cancer/Cancer genomics"},{"id":46630003,"name":"Biological sciences/Computational biology and bioinformatics/Genome informatics"}],"tags":[],"updatedAt":"2025-11-18T08:07:27+00:00","versionOfRecord":{"articleIdentity":"rs-6304422","link":"https://doi.org/10.1038/s42003-025-08994-w","journal":{"identity":"communications-biology","isVorOnly":false,"title":"Communications Biology"},"publishedOn":"2025-11-17 05:00:00","publishedOnDateReadable":"November 17th, 2025"},"versionCreatedAt":"2025-04-24 10:08:11","video":"","vorDoi":"10.1038/s42003-025-08994-w","vorDoiUrl":"https://doi.org/10.1038/s42003-025-08994-w","workflowStages":[]},"version":"v1","identity":"rs-6304422","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6304422","identity":"rs-6304422","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}