Empirical insights into the stochasticity of small RNA sequencing

Statistical Assessment of Depth Normalization for Small RNA Sequencing

PURPOSE

Methods for depth normalization have been assessed primarily with simulated data or cell-line–mixture data. There is a pressing need for benchmark data enabling a more realistic and objective assessment, especially in the context of small RNA sequencing.

METHODS

We collected a unique pair of microRNA sequencing data sets for the same set of tumor samples; one data set was collected with and the other without uniform handling and balanced design. The former provided a benchmark for evaluating evidence of differential expression and the latter served as a test bed for normalization. Next, we developed a data perturbation algorithm to simulate additional data set pairs. Last, we assembled a set of computational tools to visualize and quantify the assessment.

RESULTS

We validated the quality of the benchmark data and showed the need for normalization of the test data. For illustration, we applied the data and tools to assess the performance of 9 existing normalization methods. Among them, trimmed mean of M-values was a better scaling method, whereas the median and the upper quartiles were consistently the worst performers; one variation of remove unwanted variation had the best chance of capturing true positives but at the cost of increased false positives. In general, these methods were, at best, moderately helpful when the level of differential expression was extensive and asymmetric.

CONCLUSION

Our study (1) provides the much-needed benchmark data and computational tools for assessing depth normalization, (2) shows the dependence of normalization performance on the underlying pattern of differential expression, and (3) calls for continued research efforts to develop more effective normalization methods.

RNA-Based Therapeutics: From Antisense Oligonucleotides to miRNAs

The first therapeutic nucleic acid, a DNA oligonucleotide, was approved for clinical use in 1998. Twenty years later, in 2018, the first therapeutic RNA-based oligonucleotide was United States Food and Drug Administration (FDA) approved. This promises to be a rapidly expanding market, as many emerging biopharmaceutical companies are developing RNA interference (RNAi)-based, and RNA-based antisense oligonucleotide therapies. However, miRNA therapeutics are noticeably absent. miRNAs are regulatory RNAs that regulate gene expression. In disease states, the expression of many miRNAs is measurably altered. The potential of miRNAs as therapies and therapeutic targets has long been discussed and in the context of a wide variety of infections and diseases. Despite the great number of studies identifying miRNAs as potential therapeutic targets, only a handful of miRNA-targeting drugs (mimics or inhibitors) have entered clinical trials. In this review, we will discuss whether the investment in finding potential miRNA therapeutic targets has yielded feasible and practicable results, the benefits and obstacles of miRNAs as therapeutic targets, and the potential future of the field.

Evaluation of commercially available small RNASeq library preparation kits using low input RNA

Background

Evolving interest in comprehensively profiling the full range of small RNAs present in small tissue biopsies and in circulating biofluids, and how the profile differs with disease, has launched small RNA sequencing (RNASeq) into more frequent use. However, known biases associated with small RNASeq, compounded by low RNA inputs, have been both a significant concern and a hurdle to widespread adoption. As RNASeq is becoming a viable choice for the discovery of small RNAs in low input samples and more labs are employing it, there should be benchmark datasets to test and evaluate the performance of new sequencing protocols and operators. In a recent publication from the National Institute of Standards and Technology, Pine et al., 2018, the investigators used a commercially available set of three tissues and tested performance across labs and platforms.

Results

In this paper, we further tested the performance of low RNA input in three commonly used and commercially available RNASeq library preparation kits; NEB Next, NEXTFlex, and TruSeq small RNA library preparation. We evaluated the performance of the kits at two different sites, using three different tissues (brain, liver, and placenta) with high (1 μg) and low RNA (10 ng) input from tissue samples, or 5.0, 3.0, 2.0, 1.0, 0.5, and 0.2 ml starting volumes of plasma. As there has been a lack of robust validation platforms for differentially expressed miRNAs, we also compared low input RNASeq data with their expression profiles on three different platforms (Abcam Fireplex, HTG EdgeSeq, and Qiagen miRNome).

Conclusions

The concordance of RNASeq results on these three platforms was dependent on the RNA expression level; the higher the expression, the better the reproducibility. The results provide an extensive analysis of small RNASeq kit performance using low RNA input, and replication of these data on three downstream technologies.

Electronic supplementary material

The online version of this article (10.1186/s12864-018-4726-6) contains supplementary material, which is available to authorized users.

Modeling bias and variation in the stochastic processes of small RNA sequencing

The use of RNA-seq as the preferred method for the discovery and validation of small RNA biomarkers has been hindered by high quantitative variability and biased sequence counts. In this paper we develop a statistical model for sequence counts that accounts for ligase bias and stochastic variation in sequence counts. This model implies a linear quadratic relation between the mean and variance of sequence counts. Using a large number of sequencing datasets, we demonstrate how one can use the generalized additive models for location, scale and shape (GAMLSS) distributional regression framework to calculate and apply empirical correction factors for ligase bias. Bias correction could remove more than 40% of the bias for miRNAs. Empirical bias correction factors appear to be nearly constant over at least one and up to four orders of magnitude of total RNA input and independent of sample composition. Using synthetic mixes of known composition, we show that the GAMLSS approach can analyze differential expression with greater accuracy, higher sensitivity and specificity than six existing algorithms (DESeq2, edgeR, EBSeq, limma, DSS, voom) for the analysis of small RNA-seq data.

microRNA-guided diagnostics in clinical samples

miRNA-guided diagnostics is a powerful molecular approach for evaluating clinical samples through miRNA detection and/or visualization. To date, this approach has been successfully used to diagnose, manage, and/or monitor a wide range of neoplastic and non-neoplastic diseases. Despite the promise of miRNA-guided diagnostics, particularly in the field of minimally invasive biomarkers, several knowledge and practical issues confound or hinder translation into routine clinical practice including: miRNA sequence database errors, suboptimal RNA extraction methods, detection assay variability, a vast array of online resources for bioinformatic analyses, and non-standardized statistical analyses for miRNA clinical testing. In this review, we raise awareness of these issues and recommend research directions to help specialists in endocrinology and metabolism integrate miRNA testing into clinical decision-making.