Systematic assessment of commercially available low-input miRNA library preparation kits

Untacking small RNA profiling and RNA fragment footprinting: Approaches and challenges in library construction

Small RNAs (sRNAs) with sizes ranging from 15 to 50 nucleotides (nt) are critical regulators of gene expression control. Prior studies have shown that sRNAs are involved in a broad range of biological processes, such as organ development, tumorigenesis, and epigenomic regulation; however, emerging evidence unveils a hidden layer of diversity and complexity of endogenously encoded sRNAs profile in eukaryotic organisms, including novel types of sRNAs and the previously unknown post-transcriptional RNA modifications. This underscores the importance for accurate, unbiased detection of sRNAs in various cellular contexts. A multitude of high-throughput methods based on next-generation sequencing (NGS) are developed to decipher the sRNA expression and their modifications. Nonetheless, distinct from mRNA sequencing, the data from sRNA sequencing suffer frequent inconsistencies and high variations emanating from the adapter contaminations and RNA modifications, which overall skew the sRNA libraries. Here, we summarize the sRNA-sequencing approaches, and discuss the considerations and challenges for the strategies and methods of sRNA library construction. The pros and cons of sRNA sequencing have significant implications for implementing RNA fragment footprinting approaches, including CLIP-seq and Ribo-seq. We envision that this review can inspire novel improvements in small RNA sequencing and RNA fragment footprinting in future. This article is categorized under: RNA Evolution and Genomics > Computational Analyses of RNA RNA Processing > Processing of Small RNAs Regulatory RNAs/RNAi/Riboswitches > Biogenesis of Effector Small RNAs.

In search of the ratio of miRNA expression as robust biomarkers for constructing stable diagnostic models among multi-center data

MicroRNAs (miRNAs) are promising biomarkers for the early detection of disease, and many miRNA-based diagnostic models have been constructed to distinguish patients and healthy individuals. To thoroughly utilize the miRNA-profiling data across different sequencing platforms or multiple centers, the models accounting the batch effects were demanded for the generalization of medical application. We conducted transcription factor (TF)-mediated miRNA-miRNA interaction network analysis and adopted the within-sample expression ratios of miRNA pairs as predictive markers. The ratio of the expression values between each miRNA pair turned out to be stable across multiple data sources. A genetic algorithm-based classifier was constructed to quantify risk scores of the probability of disease and discriminate disease states from normal states in discovery, with a validation dataset for COVID-19, renal cell carcinoma, and lung adenocarcinoma. The predictive models based on the expression ratio of interacting miRNA pairs demonstrated good performances in the discovery and validation datasets, and the classifier may be used accurately for the early detection of disease.

Enhanced Detection of Novel Low-Frequency Gene Fusions via High-Yield Ligation and Multiplexed Enrichment Sequencing

Panel-based methods are commonly employed for the analysis of novel gene fusions in precision diagnostics and new drug development in cancer. However, these methods are constrained by limitations in ligation yield and the enrichment of novel gene fusions with low variant allele frequencies. In this study, we conducted a pioneering investigation into the stability of double-stranded adapter DNA, resulting in improved ligation yield and enhanced conversion efficiency. Additionally, we implemented blocker displacement amplification, achieving a remarkable 7-fold enrichment of novel gene fusions. Leveraging the pre-enrichment achieved with this approach, we successfully applied it to Nanopore sequencing, enabling ultra-fast analysis of novel gene fusions within one hour with high sensitivity. This method offers a robust and remarkably sensitive mean of analyzing novel gene fusions, promising the discovery of pivotal biomarkers that can significantly improve cancer diagnostics and the development of new therapeutic strategies.

Serum microRNA Profiles and Pathways in Hepatitis B-Associated Hepatocellular Carcinoma: A South African Study

The incidence and mortality of hepatocellular carcinoma (HCC) in Sub-Saharan Africa is projected to increase sharply by 2040 against a backdrop of limited diagnostic and therapeutic options. Two large South African-based case control studies have developed a serum-based miRNome for Hepatitis B-associated hepatocellular carcinoma (HBV-HCC), as well as identifying their gene targets and pathways. Using a combination of RNA sequencing, differential analysis and filters including a unique molecular index count (UMI) ≥ 10 and log fold change (LFC) range > 2: <-0.5 (p < 0.05), 91 dysregulated miRNAs were characterized including 30 that were upregulated and 61 were downregulated. KEGG analysis, a literature review and other bioinformatic tools identified the targeted genes and HBV-HCC pathways of the top 10 most dysregulated miRNAs. The results, which are based on differentiating miRNA expression of cases versus controls, also develop a serum-based miRNA diagnostic panel that indicates 95.9% sensitivity, 91.0% specificity and a Youden Index of 0.869. In conclusion, the results develop a comprehensive African HBV-HCC miRNome that potentially can contribute to RNA-based diagnostic and therapeutic options.

IsoSeek for unbiased and UMI-informed sequencing of miRNAs from low input samples at single-nucleotide resolution

Besides canonical microRNAs (miRNAs), sequence-based variations called isomiRs have biological relevance and diagnostic potential; however, accurate calling of these post-transcriptional modifications is challenging, especially for low input samples. Here, we present IsoSeek, a sequencing protocol that reduces ligation and PCR amplification bias and improves the accuracy of miRNA detection in low input samples. We describe steps for using randomized adapters combined with unique molecular identifiers (UMI), library quantification, and sequencing, followed by detailed procedures for data processing and analysis. For complete details on the use and execution of this protocol, please refer to C. Gómez-Martín et al. (2023)1 and Van Eijndhoven et al. (2021).2.

GABA regulates the proliferation and apoptosis of head and neck squamous cell carcinoma cells by promoting the expression of CCND2 and BCL2L1

Gamma-aminobutyric acid (GABA) is a multifunctional molecule that is widely present in the nervous system and nonneuronal tissues. It plays pivotal roles in neurotransmission, regulation of secretion, cell differentiation, proliferation, and tumorigenesis. However, the exact mechanisms of GABA in head and neck squamous cell carcinomas (HNSCCs) are unknown. We took advantage of RNA sequencing in this work and uncovered the potential gene expression profiles of the GABA-treated HNSCC cell line HN4-2. We found that the expression of CCND2 and BCL2L1 was significantly upregulated. Furthermore, GABA treatment inhibited the cell apoptosis induced by cisplatin and regulated the cell cycle after treatment with cisplatin in HN4-2 cells. Moreover, we also found that GABA could upregulate the expression of CCND2 and BCL2L1 after treatment with cisplatin. Our results not only reveal the potential pro-tumorigenic effect of GABA on HNSCCs but also provide a novel therapeutic target for HNSCC treatment.

Reassessment of miRNA variant (isomiRs) composition by small RNA sequencing

IsomiRs, sequence variants of mature microRNAs, are usually detected and quantified using high-throughput sequencing. Many examples of their biological relevance have been reported, but sequencing artifacts identified as artificial variants might bias biological inference and therefore need to be ideally avoided. We conducted a comprehensive evaluation of 10 different small RNA sequencing protocols, exploring both a theoretically isomiR-free pool of synthetic miRNAs and HEK293T cells. We calculated that, with the exception of two protocols, less than 5% of miRNA reads can be attributed to library preparation artifacts. Randomized-end adapter protocols showed superior accuracy, with 40% of true biological isomiRs. Nevertheless, we demonstrate concordance across protocols for selected miRNAs in non-templated uridyl additions. Notably, NTA-U calling and isomiR target prediction can be inaccurate when using protocols with poor single-nucleotide resolution. Our results highlight the relevance of protocol choice for biological isomiRs detection and annotation, which has key potential implications for biomedical applications.

Circulating micro-RNAs Differentially Expressed in Korean Alzheimer's Patients With Brain Aβ Accumulation Activate Amyloidogenesis

BACKGROUND

Roles for extracellular vesicles (EVs) enriched with micro-RNAs (miRNAs) have been proposed in Alzheimer's disease (AD) pathogenesis, leading to the discovery of blood miRNAs as AD biomarkers. However, the diagnostic utility of specific miRNAs is not consistent. This study aimed to discover blood miRNAs that are differentially expressed in Korean AD patients, evaluate their clinical performance, and investigate their role in amyloidogenesis.

METHODS

We discovered miRNAs differentially expressed in AD (N = 8) from cognitively normal participants (CN, N = 7) or Parkinson's disease (PD) patients (N = 8). We evaluated the clinical performance of these miRNAs in plasma of subgroup (N = 99) and in plasma EVs isolated from the total cohort (N = 251). The effects of miRNAs on amyloidogenesis and on the regulation of their target genes were investigated in vitro.

RESULTS

Among 17 upregulated and one downregulated miRNAs in AD (>twofold), miR-122-5p, miR-210-3p, and miR-590-5p were differentially expressed compared with CN or PD. However, the diagnostic performance of the selected plasma or EV miRNAs in total participants were limited (area under the curve < 0.8). Nevertheless, levels of 3 miRNAs in plasma or plasma EVs of participants who were amyloid positron emission tomography (Aβ-PET) positive were significantly higher than those from the Aβ-PET negative participants (p < .05). The selected miRNAs induced Aβ production (p < .05) through activation of β-cleavage of amyloid precursor protein (CTF-β; p < .01), and downregulated their target genes (ADAM metallopeptidase domain 10, Brain-derived neurotrophic factor, and Jagged canonical notch ligand 1; p < .05), which was further supported by pathway enrichment analysis of target genes of the miRNAs.

CONCLUSION

In conclusion, despite of the limited diagnostic utility of selected miRNAs as plasma or plasma EV biomarkers, the discovered miRNAs may play a role in amyloidogenesis during AD onset and progression.

APE1 controls DICER1 expression in NSCLC through miR-33a and miR-130b

Increasing evidence suggests different, not completely understood roles of microRNA biogenesis in the development and progression of lung cancer. The overexpression of the DNA repair protein apurinic/apyrimidinic endodeoxyribonuclease 1 (APE1) is an important cause of poor chemotherapeutic response in lung cancer and its involvement in onco-miRNAs biogenesis has been recently described. Whether APE1 regulates miRNAs acting as prognostic biomarkers of lung cancer has not been investigated, yet. In this study, we analyzed miRNAs differential expression upon APE1 depletion in the A549 lung cancer cell line using high-throughput methods. We defined a signature of 13 miRNAs that strongly correlate with APE1 expression in human lung cancer: miR-1246, miR-4488, miR-24, miR-183, miR-660, miR-130b, miR-543, miR-200c, miR-376c, miR-218, miR-146a, miR-92b and miR-33a. Functional enrichment analysis of this signature revealed its biological relevance in cancer cell proliferation and survival. We validated DICER1 as a direct functional target of the APE1-regulated miRNA-33a-5p and miR-130b-3p. Importantly, IHC analyses of different human tumors confirmed a negative correlation existing between APE1 and Dicer1 protein levels. DICER1 downregulation represents a prognostic marker of cancer development but the mechanisms at the basis of this phenomenon are still completely unknown. Our findings, suggesting that APE1 modulates DICER1 expression via miR-33a and miR-130b, reveal new mechanistic insights on DICER1 regulation, which are of relevance in lung cancer chemoresistance and cancer invasiveness.

A Unified Computational Framework for a Robust, Reliable, and Reproducible Identification of Novel miRNAs From the RNA Sequencing Data

In eukaryotic cells, miRNAs regulate a plethora of cellular functionalities ranging from cellular metabolisms, and development to the regulation of biological networks and pathways, both under homeostatic and pathological states like cancer.Despite their immense importance as key regulators of cellular processes, accurate and reliable estimation of miRNAs using Next Generation Sequencing is challenging, largely due to the limited availability of robust computational tools/methods/pipelines. Here, we introduce miRPipe, an end-to-end computational framework for the identification, characterization, and expression estimation of small RNAs, including the known and novel miRNAs and previously annotated pi-RNAs from small-RNA sequencing profiles. Our workflow detects unique novel miRNAs by incorporating the sequence information of seed and non-seed regions, concomitant with clustering analysis. This approach allows reliable and reproducible detection of unique novel miRNAs and functionally same miRNAs (paralogues). We validated the performance of miRPipe with the available state-of-the-art pipelines using both synthetic datasets generated using the newly developed miRSim tool and three cancer datasets (Chronic Lymphocytic Leukemia, Lung cancer, and breast cancer). In the experiment over the synthetic dataset, miRPipe is observed to outperform the existing state-of-the-art pipelines (accuracy: 95.23% and F1-score: 94.17%). Analysis on all the three cancer datasets shows that miRPipe is able to extract more number of known dysregulated miRNAs or piRNAs from the datasets as compared to the existing pipelines.

Development of Alzheimer’s Disease Biomarkers: From CSF- to Blood-Based Biomarkers

In the 115 years since the discovery of Alzheimer’s disease (AD), our knowledge, diagnosis, and therapeutics have significantly improved. Biomarkers are the primary tools for clinical research, diagnostics, and therapeutic monitoring in clinical trials. They provide much insightful information, and while they are not clinically used routinely, they help us to understand the mechanisms of this disease. This review charts the journey of AD biomarker discovery and development from cerebrospinal fluid (CSF) amyloid-beta 1-42 (Aβ42), total tau (T-tau), and phosphorylated tau (p-tau) biomarkers and imaging technologies to the next generation of biomarkers. We also discuss advanced high-sensitivity assay platforms for CSF Aβ42, T-tau, p-tau, and blood analysis. The recently proposed Aβ deposition/tau biomarker/neurodegeneration or neuronal injury (ATN) scheme might facilitate the definition of the biological status underpinning AD and offer a common language among researchers across biochemical biomarkers and imaging. Moreover, we highlight blood-based biomarkers for AD that offer a scalable alternative to CSF biomarkers through cost-saving and reduced invasiveness, and may provide an understanding of disease initiation and development. We discuss different groups of blood-based biomarker candidates, their advantages and limitations, and paths forward, from identification and analysis to clinical validation. The development of valid blood-based biomarkers may facilitate the implementation of future AD therapeutics and diagnostics.

Small RNA-Sequencing for Analysis of Circulating miRNAs: Benchmark Study

Small RNA-sequencing (RNA-Seq) is being increasingly used for profiling of circulating microRNAs (miRNAs), a new group of promising biomarkers. Unfortunately, small RNA-Seq protocols are prone to biases limiting quantification accuracy, which motivated development of several novel methods. Here, we present comparison of all small RNA-Seq library preparation approaches that are commercially available for quantification of miRNAs in biofluids. Using synthetic and human plasma samples, we compared performance of traditional two-adaptor ligation protocols (Lexogen, Norgen), as well as methods using randomized adaptors (NEXTflex), polyadenylation (SMARTer), circularization (RealSeq), capture probes (EdgeSeq), or unique molecular identifiers (QIAseq). There was no single protocol outperforming others across all metrics. Limited overlap of measured miRNA profiles was documented between methods largely owing to protocol-specific biases. Methods designed to minimize bias largely differ in their performance, and contributing factors were identified. Usage of unique molecular identifiers has rather negligible effect and, if designed incorrectly, can even introduce spurious results. Together, these results identify strengths and weaknesses of all current methods and provide guidelines for applications of small RNA-Seq in biomarker research.

Deciphering specific miRNAs in brain tumors: a 5-miRNA signature in glioblastoma

MicroRNAs are endogenous non-coding RNAs with a marked impact on the development and progression of brain tumors. However, they commonly share different expression patterns in other types of tumors, thereby exhibiting lack of tissue specificity. Here, an integrative holistic analysis of microarray data is established for deciphering dysregulated miRNAs in glioblastoma, distinguishing them from eight other CNS tumors. The identification of dysregulated miRNAs was performed in a pool of 176 patients, 118 of which diagnosed with glioblastoma. Dysregulated miRNAs commonly expressed in glioblastoma were then discriminated from those co-expressed in other CNS tumors and further characterized. Overall, 21 miRNAs were found to be commonly dysregulated in glioblastoma. Notwithstanding, 16 miRNAs also exhibited a differential expression in at least one other CNS tumor. The remaining 5, specifically, hsa-miR-21-3p, hsa-miR-338-5p, hsa-miR-485-5p, hsa-miR-491-5p and hsa-miR-1290, were solely associated to glioblastoma. This signature is in-depth characterized, with the spotlight on tumor progression, invasion and patient survival. These five endogenous molecules, differentially expressed in glioblastoma, are thus suggested as potential therapeutic targets, modulating several genes involved in major signalling pathways, including MAPK/ERK, calcium, PI3K/AKT, mTOR and Wnt. In summary, these findings lay a foundation for further research on the expression and function of specific patterns of miRNAs expression in glioblastoma, providing reference for potential novel targets.

Non-Coding RNAs in Preeclampsia-Molecular Mechanisms and Diagnostic Potential

Preeclampsia (PE) is a leading cause of maternal and neonatal morbidity and mortality worldwide. Defects in trophoblast invasion, differentiation of extravillous trophoblasts and spiral artery remodeling are key factors in PE development. Currently there are no predictive biomarkers clinically available for PE. Recent technological advancements empowered transcriptome exploration and led to the discovery of numerous non-coding RNA species of which microRNAs (miRNAs) and long non-coding RNAs (lncRNAs) are the most investigated. They are implicated in the regulation of numerous cellular functions, and as such are being extensively explored as potential biomarkers for various diseases. Altered expression of numerous lncRNAs and miRNAs in placenta has been related to pathophysiological processes that occur in preeclampsia. In the following text we offer summary of the latest knowledge of the molecular mechanism by which lnRNAs and miRNAs (focusing on the chromosome 19 miRNA cluster (C19MC)) contribute to pathophysiology of PE development and their potential utility as biomarkers of PE, with special focus on sample selection and techniques for the quantification of lncRNAs and miRNAs in maternal circulation.

Chemical capping improves template switching and enhances sequencing of small RNAs

Template-switching reverse transcription is widely used in RNA sequencing for low-input and low-quality samples, including RNA from single cells or formalin-fixed paraffin-embedded (FFPE) tissues. Previously, we identified the native eukaryotic mRNA 5′ cap as a key structural element for enhancing template switching efficiency. Here, we introduce CapTS-seq, a new strategy for sequencing small RNAs that combines chemical capping and template switching. We probed a variety of non-native synthetic cap structures and found that an unmethylated guanosine triphosphate cap led to the lowest bias and highest efficiency for template switching. Through cross-examination of different nucleotides at the cap position, our data provided unequivocal evidence that the 5′ cap acts as a template for the first nucleotide in reverse transcriptase-mediated post-templated addition to the emerging cDNA—a key feature to propel template switching. We deployed CapTS-seq for sequencing synthetic miRNAs, human total brain and liver FFPE RNA, and demonstrated that it consistently improves library quality for miRNAs in comparison with a gold standard template switching-based small RNA-seq kit.

Single-cell microRNA sequencing method comparison and application to cell lines and circulating lung tumor cells

Molecular single cell analyses provide insights into physiological and pathological processes. Here, in a stepwise approach, we first evaluate 19 protocols for single cell small RNA sequencing on MCF7 cells spiked with 1 pg of 1,006 miRNAs. Second, we analyze MCF7 single cell equivalents of the eight best protocols. Third, we sequence single cells from eight different cell lines and 67 circulating tumor cells (CTCs) from seven SCLC patients. Altogether, we analyze 244 different samples. We observe high reproducibility within protocols and reads covered a broad spectrum of RNAs. For the 67 CTCs, we detect a median of 68 miRNAs, with 10 miRNAs being expressed in 90% of tested cells. Enrichment analysis suggested the lung as the most likely organ of origin and enrichment of cancer-related categories. Even the identification of non-annotated candidate miRNAs was feasible, underlining the potential of single cell small RNA sequencing.

Technologies for small non-coding RNA sequencing at the single-cell level are less mature than for sequencing mRNAs. Here the authors evaluate available protocols for analysis of circulating lung cancer tumour cells.

Protocol to analyze circulating small non-coding RNAs by high-throughput RNA sequencing from human plasma samples

The identification and validation of circulating small non-coding RNA (sncRNA) as biomarkers for disease diagnosis, staging, and response to novel therapies is still a compelling challenge. Pre-analytical variables, such as storage temperature or blood hemolysis, and different analytical approaches affect sncRNA stability, detection, and expression, resulting in discrepancies among studies. Here, we report a systematic standardized protocol to reproducibly analyze circulating sncRNAs, employing high-throughput sncRNA sequencing and qRT-PCR validation, from 200 μL of human plasma samples. For details on the use and execution of this protocol, please refer to Ventriglia et al. (2020), Sebastiani et al. (2017), and Dotta et al. (2018).

Small RNA-Sequencing: Approaches and Considerations for miRNA Analysis

MicroRNAs (miRNAs) are a class of small RNA molecules that have an important regulatory role in multiple physiological and pathological processes. Their disease-specific profiles and presence in biofluids are properties that enable miRNAs to be employed as non-invasive biomarkers. In the past decades, several methods have been developed for miRNA analysis, including small RNA sequencing (RNA-seq). Small RNA-seq enables genome-wide profiling and analysis of known, as well as novel, miRNA variants. Moreover, its high sensitivity allows for profiling of low input samples such as liquid biopsies, which have now found applications in diagnostics and prognostics. Still, due to technical bias and the limited ability to capture the true miRNA representation, its potential remains unfulfilled. The introduction of many new small RNA-seq approaches that tried to minimize this bias, has led to the existence of the many small RNA-seq protocols seen today. Here, we review all current approaches to cDNA library construction used during the small RNA-seq workflow, with particular focus on their implementation in commercially available protocols. We provide an overview of each protocol and discuss their applicability. We also review recent benchmarking studies comparing each protocol's performance and summarize the major conclusions that can be gathered from their usage. The result documents variable performance of the protocols and highlights their different applications in miRNA research. Taken together, our review provides a comprehensive overview of all the current small RNA-seq approaches, summarizes their strengths and weaknesses, and provides guidelines for their applications in miRNA research.

Quantitative mapping of the cellular small RNA landscape with AQRNA-seq

Current next-generation RNA sequencing methods do not provide accurate quantification of small RNAs within a sample due to sequence-dependent biases in capture, ligation, and amplification during library preparation. We present a method, Absolute Quantification (AQ) RNA-seq, that minimizes biases and provides a direct, linear correlation between sequencing read count and copy number for all small RNAs in a sample. Library preparation and data processing were optimized and validated using a 963-member miRNA reference library, oligonucleotide standards of varying lengths, and northern blots. Application of AQRNA-seq to a panel of human cancer cells revealed >800 detectable miRNAs that varied during cancer progression, while application to bacterial tRNA pools, with the challenges of secondary structure and abundant modifications, revealed 80-fold variation in tRNA isoacceptor levels, stress-induced site-specific tRNA fragmentation, quantitative modification maps, and evidence for stress-induced tRNA-driven codon-biased translation. AQRNA-seq thus provides a versatile means to quantitatively map the small RNA landscape in cells.

CoolMPS: evaluation of antibody labeling based massively parallel non-coding RNA sequencing

Results of massive parallel sequencing-by-synthesis vary depending on the sequencing approach. CoolMPS™ is a new sequencing chemistry that incorporates bases by labeled antibodies. To evaluate the performance, we sequenced 240 human non-coding RNA samples (dementia patients and controls) with and without CoolMPS. The Q30 value as indicator of the per base sequencing quality increased from 91.8 to 94%. The higher quality was reached across the whole read length. Likewise, the percentage of reads mapping to the human genome increased from 84.9 to 86.2%. For both technologies, we computed similar distributions between different RNA classes (miRNA, piRNA, tRNA, snoRNA and yRNA) and within the classes. While standard sequencing-by-synthesis allowed to recover more annotated miRNAs, CoolMPS yielded more novel miRNAs. The correlation between the two methods was 0.97. Evaluating the diagnostic performance, we observed lower minimal P-values for CoolMPS (adjusted P-value of 0.0006 versus 0.0004) and larger effect sizes (Cohen's d of 0.878 versus 0.9). Validating 19 miRNAs resulted in a correlation of 0.852 between CoolMPS and reverse transcriptase-quantitative polymerase chain reaction. Comparison to data generated with Illumina technology confirmed a known shift in the overall RNA composition. With CoolMPS we evaluated a novel sequencing-by-synthesis technology showing high performance for the analysis of non-coding RNAs.

A low-bias and sensitive small RNA library preparation method using randomized splint ligation

Small RNAs are important regulators of gene expression and are involved in human development and disease. Next generation sequencing (NGS) allows for scalable, genome-wide studies of small RNA; however, current methods are challenged by low sensitivity and high bias, limiting their ability to capture an accurate representation of the cellular small RNA population. Several studies have shown that this bias primarily arises during the ligation of single-strand adapters during library preparation, and that this ligation bias is magnified by 2′-O-methyl modifications (2′OMe) on the 3′ terminal nucleotide. In this study, we developed a novel library preparation process using randomized splint ligation with a cleavable adapter, a design which resolves previous challenges associated with this ligation strategy. We show that a randomized splint ligation based workflow can reduce bias and increase the sensitivity of small RNA sequencing for a wide variety of small RNAs, including microRNA (miRNA) and tRNA fragments as well as 2′OMe modified RNA, including Piwi-interacting RNA and plant miRNA. Finally, we demonstrate that this workflow detects more differentially expressed miRNA between tumorous and matched normal tissues. Overall, this library preparation process allows for highly accurate small RNA sequencing and will enable studies of 2′OMe modified RNA with new levels of detail.

Insufficiently complex unique-molecular identifiers (UMIs) distort small RNA sequencing

The attachment of unique molecular identifiers (UMIs) to RNA molecules prior to PCR amplification and sequencing, makes it possible to amplify libraries to a level that is sufficient to identify rare molecules, whilst simultaneously eliminating PCR bias through the identification of duplicated reads. Accurate de-duplication is dependent upon a sufficiently complex pool of UMIs to allow unique labelling. In applications dealing with complex libraries, such as total RNA-seq, only a limited variety of UMIs are required as the variation in molecules to be sequenced is enormous. However, when sequencing a less complex library, such as small RNAs for which there is a more limited range of possible sequences, we find increased variation in UMIs are required, even beyond that provided in a commercial kit specifically designed for the preparation of small RNA libraries for sequencing. We show that a pool of UMIs randomly varying across eight nucleotides is not of sufficient depth to uniquely tag the microRNAs to be sequenced. This results in over de-duplication of reads and the marked under-estimation of expression of the more abundant microRNAs. Whilst still arguing for the utility of UMIs, this work demonstrates the importance of their considered design to avoid errors in the estimation of gene expression in libraries derived from select regions of the transcriptome or small genomes.

Identification of a potential non-coding RNA biomarker signature for amyotrophic lateral sclerosis

Objective biomarkers for the clinically heterogeneous adult-onset neurodegenerative disorder amyotrophic lateral sclerosis are crucial to facilitate assessing emerging therapeutics and improve the diagnostic pathway in what is a clinically heterogeneous syndrome. With non-coding RNA transcripts including microRNA, piwi-RNA and transfer RNA present in human biofluids, we sought to identify whether non-coding RNA in serum could be biomarkers for amyotrophic lateral sclerosis. Serum samples from our Oxford Study for Biomarkers in motor neurone disease/amyotrophic lateral sclerosis discovery cohort of amyotrophic lateral sclerosis patients (n = 48), disease mimics (n = 16) and age- and sex-matched healthy controls (n = 24) were profiled for non-coding RNA expression using RNA-sequencing, which showed a wide range of non-coding RNA to be dysregulated. We confirmed significant alterations with reverse transcription-quantitative PCR in the expression of hsa-miR-16-5p, hsa-miR-21-5p, hsa-miR-92a-3p, hsa-piR-33151, TRV-AAC4-1.1 and TRA-AGC6-1.1. Furthermore, hsa-miR-206, a previously identified amyotrophic lateral sclerosis biomarker, showed a binary-like pattern of expression in our samples. Using the expression of these non-coding RNA, we were able to discriminate amyotrophic lateral sclerosis samples from healthy controls in our discovery cohort using a random forest analysis with 93.7% accuracy with promise in predicting progression rate of patients. Importantly, cross-validation of this novel signature using a new geographically distinct cohort of samples from the United Kingdom and Germany with both amyotrophic lateral sclerosis and control samples (n = 156) yielded an accuracy of 73.9%. The high prediction accuracy of this non-coding RNA-based biomarker signature, even across heterogeneous cohorts, demonstrates the strength of our approach as a novel platform to identify and stratify amyotrophic lateral sclerosis patients.