Long-read single-cell RNA-sequencing enables the profiling of RNA isoform expression and alternative splicing at the single cell level. However, diverse single-cell technologies and sparse isoform data demand flexible, accurate, processing and analysis tools. Here we introduce FLAMESv2, a highly modular and protocol-agnostic R/Bioconductor package for long-read singlecell RNA-seq data processing. FLAMESv2 supports a wide range of single-cell and spatial experimental protocols, is highly configurable and scalable, allowing seamless multi-sample analysis and provides versatile visualisation and analysis outputs. We demonstrate FLAMESv2 compatibility with both dropletbased and combinatorial barcoding single-cell methods, as well as spatial transcriptomics workflows. Applying FLAMESv2 to in vitro differentiation of stem cells into neurons, we identify celltypes, differentiation trajectories, expression of annotated and novel isoforms and isoform expression diversity and heterogeneity within individual cells. FLAMESv2 provides a comprehensive, flexible approach to analysing long-read single-cell RNA-sequencing, unlocking this powerful methodology for RNA isoform characterisation. ### Competing Interest Statement Y.D.J.P., R.D.P., A.L., Q.G., N.M.D., M.B.C. and Y.Y. have received support from Oxford Nanopore Technologies (ONT) to present their findings at scientific conferences. However, ONT played no role in the study design, execution, analysis, or publication of this research. Chan Zuckerberg Initiative DAF, 2019-002443 Australian National Health and Medical Research Council, GNT2007996, GNT2016547, GNT1196841, GNT2017257 Australian Research Council, 200102460, 200102903 Victorian State Government Operational Infrastructure Support and Australian Government, 11F9623N, 11F9625N

BACKGROUND:Hypoxic microenvironment is a common feature of most solid tumors including hepatocellular carcinoma (HCC). Vasculogenic mimicry (VM) formation by tumor cells could provide blood supply to tumor cells under hypoxia. NFE2 like basic leucine zipper (bZIP) transcription factor 2 (Nrf2), a regulator of cellular homeostasis, may promote tumor progression in the hypoxic conditions. However, the role and regulatory mechanisms of Nrf2 in HCC are not fully elucidated. METHODS:Nrf2 and assembly factor for spindle microtubules (ASPM) expression modulations were conducted by lentiviral transfections. Western blot, immunofluorescence, ChIP-qPCR, dual-luciferase reporter gene assay, flow cytometry, RNA sequencing, multiple bioinformatics databases analysis, cell function assays in vitro, mouse model in vivo and human HCC tissues were employed to assess the effect of Nrf2/ASPM axis on HCC progression under hypoxia. RESULTS:Nrf2 and ASPM expression facilitated epithelial-mesenchymal transition (EMT), cancer stem cells (CSCs) feature, and VM formation of HCC cells under hypoxia. Furthermore, Nrf2-regulated ASPM expression, via binding directly to the promoter region of ASPM and transcriptionally promoting ASPM expression. ASPM re-expression in Nrf2 knockdown cells or ASPM knockdown in Nrf2 overexpression cells reversed the cellular function caused by Nrf2. Meantime, retinol metabolism pathway was disrupted following abnormal ASPM expression. Nrf2/ASPM axis in murine models accelerated tumor growth and VM, corroborating in vitro findings. All-trans retinoic acid treatment reversed stemness and VM of HCC cells in vitro and in vivo. Clinically, Nrf2 and ASPM expressions were related to poor prognosis of HCC patients. CONCLUSIONS:Nrf2 drives EMT, CSCs characteristics and VM in HCC under hypoxia through the modulation of ASPM. Retinol metabolism pathway was dysregulated in HCC cells with ASPM overexpression. Nrf2/ASPM axis and related pathway provided potential therapeutic target for HCC.

[This corrects the article on p. 260 in vol. 7, PMID: 28337375.].

A hypoxic microenvironment is the most frequent characteristic in tumor microenvironment. Programmed death-ligand 1 (PD-L1) is an important molecule and therapeutic target that mediates the immune response of tumor cells. Previous studies have shown that hypoxia can lead to increased expression of Nucleophosmin 1 (NPM1) and PD-L1. However, the exact regulatory mechanisms of NPM1 and PD-L1 expression under hypoxic conditions are still poorly understood. The relationships among hypoxia, NPM1 and PD-L1 were explored by western blotting, immunofluorescence staining, flow cytometry and chromatin immunoprecipitation-quantitativePCR(ChIP-qPCR). Animal tumor models were established to explore the effect of NPM1 expression on tumor growth. The relationships between NPM1 and breast cancer (BC) clinical features and immune infiltration were revealed by bioinformatics analysis. NPM1 mediates increased PD-L1 expression in the hypoxic microenvironment of BC. HIF-1α can increase the expression of NPM1 by activating the p-AKT pathway and binding to the NPM1 promoter. Increased expression of NPM1 can promote tumor growth and inhibit T cell infiltration. Bioinformatics analysis showed that the high expression of NPM1 was associated with poorer survival and immunosuppression in patients with BC. The hypoxic microenvironment promotes PD-L1 expression via NPM1 in BC, which may be further associated with the inhibition of tumor immunity. NPM1 may serve as a potential target for modulating PD-L1 immunotherapy.

Differential expression analysis of RNA-seq is one of the most commonly performed bioinformatics analyses. Transcript-level quantifications are inherently more uncertain than gene-level read counts because of ambiguous assignment of sequence reads to transcripts. While sequence reads can usually be assigned unambiguously to a gene, reads are very often compatible with multiple transcripts for that gene, particularly for genes with many isoforms. Software tools designed for gene-level differential expression do not perform optimally on transcript counts because the read-to-transcript ambiguity (RTA) disrupts the mean-variance relationship normally observed for gene level RNA-seq data and interferes with the efficiency of the empirical Bayes dispersion estimation procedures. The pseudoaligners kallisto and Salmon provide bootstrap samples from which quantification uncertainty can be assessed. We show that the overdispersion arising from RTA can be elegantly estimated by fitting a quasi-Poisson model to the bootstrap counts for each transcript. The technical overdispersion arising from RTA can then be divided out of the transcript counts, leading to scaled counts that can be input for analysis by established gene-level software tools with full statistical efficiency. Comprehensive simulations and test data show that an edgeR analysis of the scaled counts is more powerful and efficient than previous differential transcript expression pipelines while providing correct control of the false discovery rate. Simulations explore a wide range of scenarios including the effects of paired vs single-end reads, different read lengths and different numbers of replicates.

Abstract Background Aberrant expressions of desmoglein 2 (Dsg2) and desmocollin 2(Dsc2), the two most widely distributed desmosomal cadherins, have been found to play various roles in cancer in a context-dependent manner. Their specific roles on breast cancer (BC) and the potential mechanisms remain unclear. Methods The expressions of Dsg2 and Dsc2 in human BC tissues and cell lines were assessed by using bioinformatics analysis, immunohistochemistry and western blotting assays. Wound-healing and Transwell assays were performed to evaluate the cells’ migration and invasion abilities. Plate colony-forming and MTT assays were used to examine the cells’ capacity of proliferation. Mechanically, Dsg2 and Dsc2 knockdown-induced malignant behaviors were elucidated using western blotting assay as well as three inhibitors including MK2206 for AKT, PD98059 for ERK, and XAV-939 for β-catenin. Results We found reduced expressions of Dsg2 and Dsc2 in human BC tissues and cell lines compared to normal counterparts. Furthermore, shRNA-mediated downregulation of Dsg2 and Dsc2 could significantly enhance cell proliferation, migration and invasion in triple-negative MDA-MB-231 and luminal MCF-7 BC cells. Mechanistically, EGFR activity was decreased but downstream AKT and ERK pathways were both activated maybe through other activated protein tyrosine kinases in shDsg2 and shDsc2 MDA-MB-231 cells since protein tyrosine kinases are key drivers of triple-negative BC survival. Additionally, AKT inhibitor treatment displayed much stronger capacity to abolish shDsg2 and shDsc2 induced progression compared to ERK inhibition, which was due to feedback activation of AKT pathway induced by ERK inhibition. In contrast, all of EGFR, AKT and ERK activities were attenuated, whereas β-catenin was accumulated in shDsg2 and shDsc2 MCF-7 cells. These results indicate that EGFR-targeted therapy is not a good choice for BC patients with low Dsg2 or Dsc2 expression. Comparatively, AKT inhibitors may be more helpful to triple-negative BC patients with low Dsg2 or Dsc2 expression, while therapies targeting β-catenin can be considered for luminal BC patients with low Dsg2 or Dsc2 expression. Conclusion Our finding demonstrate that single knockdown of Dsg2 or Dsc2 could promote proliferation, motility and invasion in triple-negative MDA-MB-231 and luminal MCF-7 cells. Nevertheless, the underlying mechanisms were cellular context-specific and distinct.

Supplementary Table 5 and 6: TRA2A expression was significantly associated with poorer survival

Group heteroscedasticity is commonly observed in pseudo-bulk single-cell RNA-seq datasets and its presence can hamper the detection of differentially expressed genes. Since most bulk RNA-seq methods assume equal group variances, we introduce two new approaches that account for heteroscedastic groups, namely voomByGroup and voomWithQualityWeights using a blocked design (voomQWB). Compared to current gold-standard methods that do not account for group heteroscedasticity, we show results from simulations and various experiments that demonstrate the superior performance of voomByGroup and voomQWB in terms of error control and power when group variances in pseudo-bulk single-cell RNA-seq data are unequal.

IntroductionHypoxia is an important characteristic of solid tumors. However, spatial transcriptomics (ST) of hypoxia-associated heterogeneity is not clear.MethodsThis study integrated Spatial Transcriptomics (ST) with immunofluorescence to demonstrate their spatial distribution in human claudin-low breast cancer MDA-MB-231 engraft. ST spots were clustered with differentially expression genes. The data were combined with hypoxia-specific marker and angiogenesis marker-labeled serial sections to indicate the spatial distribution of hypoxia and hypoxia-inducted transcriptional profile. Moreover, marker genes, cluster-specific hypoxia genes, and their co-essential relationship were identified and mapped in every clusters. The clinicopathological association of marker genes of hypoxia-dependent spatial clusters was explored in 1904 breast cancers from METABRIC database.ResultsThe tumor from center to periphery were enriched into five hypoxia-dependent subgroups with differentially expressed genes, which were matched to necrosis, necrosis periphery, hypoxic tumor, adaptive survival tumor, and invasive tumor, respectively. Different subgroups demonstrated distinct hypoxia condition and spatial heterogeneity in biological behavior and signaling pathways. Cox regression analysis showed that the invasive tumor (cluster 0) and hypoxic tumor (cluster 6) score could be served as independent prognostic factors in claudin-low patients. KM analysis indicated that high invasive tumor (cluster 0) and hypoxic tumor (cluster 6) score was associated with poor prognoses of claudin-low patients. Further analysis showed that hypoxia-induced immune checkpoints, such as CD276 and NRP1, upregulation in invasive tumor to block infiltration and activation of B cells and CD8+ T cells to change tumor immune microenvironment. DiscussionThis study reveals hypoxia-dependent spatial heterogeneity in claudin-low breast cancer and highlights its potential value as a predictive biomarker of clinical outcomes and immunotherapy response. The molecules found in this study also provided potential molecular mechanisms and therapeutic targets for subsequent studies.

Supplementary Figure legends; Table S1, S4, S7, S8, and RSRC2 minigene sequence. Table S1: The increased splicing factors expression induced by PTX treatment in MDA-MB-231 cells; Table S4. All AS events selected for validation were confirmed by VeriQuest SYBR Green qPCR; Table S7: The relationship of TRA2A, CALUl, CALUs, RSRC2l, RSRC2s, PALMl, PALMs mRNA and TRA2A protein expression with PTXR of TNBC; Table S8: The relationship of TRA2A, CALUl, CALUs, RSRC2l, RSRC2s, PALMl and PALMs mRNA with TRA2A protein expression.

Triple-negative breast cancer (TNBC) has a shorter survival time and higher mortality rate than other molecular subtypes. RSRC2 is a newly discovered tumor suppressor gene. However, the potential functional mechanism of RSRC2 in TNBC remains unknown so far. Multiple bioinformatics databases were used. A Human Transcriptome Array 2.0 analysis, ChIP-seq analysis, ChIP-qPCR, RT-qPCR, Western blot, cell function assays in vitro and a metastatic mouse model in vivo were performed to demonstrate the role of RSRC2 in TNBC. Through the analysis of various databases, RSRC2 expression was the lowest in TNBC tissues compared to other molecular subtypes. The low expression of RSRC2 was associated with a worse prognosis for patients with breast cancer. The transcriptome array, ChIP-seq and bioinformatics analysis identified that GRHL2 and SCIN might have a close relationship with RSRC2. The functional bioinformatics enrichment analysis and functional cell experiments showed that RSRC2 was involved in cell adhesion, cell proliferation, cell migration and invasion. Furthermore, RSRC2 expression suppressed SCIN expression but not GRHL2 expression. SCIN re-expression in the RSRC2 overexpression cells or SCIN knockdown in the RSRC2 knockdown cells reversed the cellular function caused by RSRC2. Mechanistically, RSRC2 transcriptionally inhibited SCIN expression. In summary, our study reveals that RSRC2 acts as a tumor suppressor in TNBC development and progression through negatively regulating SCIN-mediated cell function, thus providing a potential target for TNBC treatment.

Table S2: TRA2A-regulated AS events

qRT-PCR showed that MELK, CEACAM1, LMCD1 and RFWD2 AS events weren't shown during PTX treatment.

We conducted a comprehensive long-read RNA sequencing (RNA-seq) benchmarking experiment by combining spike-ins and in silico mixtures to establish a ground-truth dataset. We used long- and short-read RNA-seq technology to deeply sequence samples and compared the performance of a range of analysis tools on these data.

The lack of benchmark data sets with inbuilt ground-truth makes it challenging to compare the performance of existing long-read isoform detection and differential expression analysis workflows. Here, we present a benchmark experiment using two human lung adenocarcinoma cell lines that were each profiled in triplicate together with synthetic, spliced, spike-in RNAs (sequins). Samples were deeply sequenced on both Illumina short-read and Oxford Nanopore Technologies long-read platforms. Alongside the ground-truth available via the sequins, we created in silico mixture samples to allow performance assessment in the absence of true positives or true negatives. Our results show that StringTie2 and bambu outperformed other tools from the six isoform detection tools tested, DESeq2, edgeR and limma-voom were best among the five differential transcript expression tools tested and there was no clear front-runner for performing differential transcript usage analysis between the five tools compared, which suggests further methods development is needed for this application.

The gene structure view of CALU (exon2 inclusion), RSRC2 (exon4 inclusion), CEACAM1 (exon7 inclusion), LMCD1 (exon3 exclusion), PALM (exon8 exclusion) and RFWD2 (exon7 exclusion).

Supplementary Figure 2. AURKA overexpression induced extra centrosomes, ALDH1 positivity and USP44 upregulation.