Bias-invariant RNA-sequencing metadata annotation.
Ontology highlight
ABSTRACT: Background
Recent technological advances have resulted in an unprecedented increase in publicly available biomedical data, yet the reuse of the data is often precluded by experimental bias and a lack of annotation depth and consistency. Missing annotations makes it impossible for researchers to find datasets specific to their needs.Findings
Here, we investigate RNA-sequencing metadata prediction based on gene expression values. We present a deep-learning-based domain adaptation algorithm for the automatic annotation of RNA-sequencing metadata. We show, in multiple experiments, that our model is better at integrating heterogeneous training data compared with existing linear regression-based approaches, resulting in improved tissue type classification. By using a model architecture similar to Siamese networks, the algorithm can learn biases from datasets with few samples.Conclusion
Using our novel domain adaptation approach, we achieved metadata annotation accuracies up to 15.7% better than a previously published method. Using the best model, we provide a list of >10,000 novel tissue and sex label annotations for 8,495 unique SRA samples. Our approach has the potential to revive idle datasets by automated annotation making them more searchable.
SUBMITTER: Wartmann H
PROVIDER: S-EPMC8559615 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Large volumes of high-throughput sequencing data have been submitted to the Sequencing Read Archive (SRA). The lack of experimental metadata associated with the data makes reuse and understanding data quality very difficult. In the case of RNA sequencing (RNA-Seq), which reveals the presence and quantity of RNA in a biological sample at any moment, it is necessary to consider that gene expression responds over a short time interval (several seconds to a few minutes) in many organisms. Therefore, to isolate RNA that accurately reflects the transcriptome at the point of harvest, raw biological samples should be processed by freezing in liquid nitrogen, immersing in RNA stabilization reagent or lysing and homogenizing in RNA lysis buffer containing guanidine thiocyanate as soon as possible. As the number of samples handled simultaneously increases, the time until the RNA is protected can increase. Here, to evaluate the effect of different lag times in RNA protection on RNA-Seq data, we harvested CHO-S cells after 3, 5, 6, and 7 days of cultivation, added RNA lysis buffer in a time course of 15, 30, 45, and 60 min after harvest, and conducted RNA-Seq. These RNA samples showed high RNA integrity number (RIN) values indicating non-degraded RNA, and sequence data from libraries prepared with these RNA samples was of high quality according to FastQC. We observed that, at the same cultivation day, global trends of gene expression were similar across the time course of addition of RNA lysis buffer; however, the expression of some genes was significantly different between the time-course samples of the same cultivation day; most of these differentially expressed genes were related to apoptosis. We conclude that the time lag between sample harvest and RNA protection influences gene expression of specific genes. It is, therefore, necessary to know not only RIN values of RNA and the quality of the sequence data but also how the experiment was performed when acquiring RNA-Seq data from the database.
| S-EPMC8401820 | biostudies-literature
Project description:Human untargeted metabolomics studies annotate only ~10% of molecular features. We introduce reference-data-driven analysis to match metabolomics tandem mass spectrometry (MS/MS) data against metadata-annotated source data as a pseudo-MS/MS reference library. Applying this approach to food source data, we show that it increases MS/MS spectral usage 5.1-fold over conventional structural MS/MS library matches and allows empirical assessment of dietary patterns from untargeted data.
| S-EPMC10277029 | biostudies-literature
Project description:RNA-seq is a powerful technique for identifying and quantifying transcription and splicing events, both known and novel. However, given its recent development and the proliferation of library construction methods, understanding the bias it introduces is incomplete but critical to realizing its value.We present a method, in vitro transcription sequencing (IVT-seq), for identifying and assessing the technical biases in RNA-seq library generation and sequencing at scale. We created a pool of over 1,000 in vitro transcribed RNAs from a full-length human cDNA library and sequenced them with polyA and total RNA-seq, the most common protocols. Because each cDNA is full length, and we show in vitro transcription is incredibly processive, each base in each transcript should be equivalently represented. However, with common RNA-seq applications and platforms, we find 50% of transcripts have more than two-fold and 10% have more than 10-fold differences in within-transcript sequence coverage. We also find greater than 6% of transcripts have regions of dramatically unpredictable sequencing coverage between samples, confounding accurate determination of their expression. We use a combination of experimental and computational approaches to show rRNA depletion is responsible for the most significant variability in coverage, and several sequence determinants also strongly influence representation.These results show the utility of IVT-seq for promoting better understanding of bias introduced by RNA-seq. We find rRNA depletion is responsible for substantial, unappreciated biases in coverage introduced during library preparation. These biases suggest exon-level expression analysis may be inadvisable, and we recommend caution when interpreting RNA-seq results.
| S-EPMC4197826 | biostudies-literature
Project description:High throughput sequencing technology provides us unprecedented opportunities to study transcriptome dynamics. Compared to microarray-based gene expression profiling, RNA-Seq has many advantages, such as high resolution, low background, and ability to identify novel transcripts. Moreover, for genes with multiple isoforms, expression of each isoform may be estimated from RNA-Seq data. Despite these advantages, recent work revealed that base level read counts from RNA-Seq data may not be randomly distributed and can be affected by local nucleotide composition. It was not clear though how the base level read count bias may affect gene level expression estimates.In this paper, by using five published RNA-Seq data sets from different biological sources and with different data preprocessing schemes, we showed that commonly used estimates of gene expression levels from RNA-Seq data, such as reads per kilobase of gene length per million reads (RPKM), are biased in terms of gene length, GC content and dinucleotide frequencies. We directly examined the biases at the gene-level, and proposed a simple generalized-additive-model based approach to correct different sources of biases simultaneously. Compared to previously proposed base level correction methods, our method reduces bias in gene-level expression estimates more effectively.Our method identifies and corrects different sources of biases in gene-level expression measures from RNA-Seq data, and provides more accurate estimates of gene expression levels from RNA-Seq. This method should prove useful in meta-analysis of gene expression levels using different platforms or experimental protocols.
| S-EPMC3149584 | biostudies-literature
Project description:BackgroundNext-generation sequencing does not yield fully unbiased estimates for read abundance, which may impact on the conclusions that can be drawn from sequencing data. The ligation step in RNA sequencing library generation is a known source of bias, motivating developments in enzyme technology and library construction protocols. We present the first comparison of the standard duplex adaptor protocol supplied by Life Technologies for use on the Ion Torrent PGM with an alternate single adaptor approach involving CircLigase (CircLig protocol).A correlation between over-representation in sequenced libraries and degree of secondary structure has been reported previously, therefore we also investigated whether bias could be reduced by ligation with an enzyme that functions at a temperature not permissive for such structure.ResultsA pool of small RNA fragments of known composition was converted into a sequencing library using one of three protocols and sequenced on an Ion Torrent PGM. The CircLig protocol resulted in less over-representation of specific sequences than the standard protocol. Over-represented sequences are more likely to be predicted to have secondary structure and to co-fold with adaptor sequences. However, use of the thermostable ligase Methanobacterium thermoautotrophicum RNA ligase K97A (Mth K97A) was not sufficient to reduce bias.ConclusionsThe single adaptor CircLigase-based approach significantly reduces, but does not eliminate, bias in Ion Torrent data. Ligases that function at temperatures to remove the possible influence of secondary structure on library generation may be of value, although Mth K97A is not effective in this case.
| S-EPMC4117970 | biostudies-literature
Project description:We found tremendous inequality across gene and protein annotation resources. We observed that this bias leads biomedical researchers to focus on richly annotated genes instead of those with the strongest molecular data. We advocate that researchers reduce these biases by pursuing data-driven hypotheses.
| S-EPMC5778030 | biostudies-literature
Project description:MicroRNAs (miRs) are short non-coding RNAs that fine-tune the regulation of gene expression to coordinate a wide range of biological processes. Because of their role in the regulation of gene expression, miRs are essential players in development by acting on cell fate determination and progression towards cell differentiation and are increasingly relevant to human health and disease. Although the zebrafish Danio rerio is a major model for studies of development, genetics, physiology, evolution, and human biology, the annotation of zebrafish miR-producing genes remains limited. In the present work, we report deep sequencing data of zebrafish small RNAs from brain, heart, testis, and ovary. Results provide evidence for the expression of 56 un-annotated mir genes and 248 un-annotated mature strands, increasing the number of zebrafish mir genes over those already deposited in miRBase by 16% and the number of mature sequences by 63%. We also describe the existence of three pairs of mirror-mir genes and two mirtron genes, genetic features previously undescribed in non-mammalian vertebrates. This report provides information that substantially increases our knowledge of the zebrafish miRNome and will benefit the entire miR community.
| S-EPMC4130647 | biostudies-literature
Project description:MotivationSingle-cell RNA sequencing (scRNA-seq) technologies enable the study of transcriptional heterogeneity at the resolution of individual cells and have an increasing impact on biomedical research. However, it is known that these methods sometimes wrongly consider two or more cells as single cells, and that a number of so-called doublets is present in the output of such experiments. Treating doublets as single cells in downstream analyses can severely bias a study's conclusions, and therefore computational strategies for the identification of doublets are needed.ResultsWith scds, we propose two new approaches for in silico doublet identification: Co-expression based doublet scoring (cxds) and binary classification based doublet scoring (bcds). The co-expression based approach, cxds, utilizes binarized (absence/presence) gene expression data and, employing a binomial model for the co-expression of pairs of genes, yields interpretable doublet annotations. bcds, on the other hand, uses a binary classification approach to discriminate artificial doublets from original data. We apply our methods and existing computational doublet identification approaches to four datasets with experimental doublet annotations and find that our methods perform at least as well as the state of the art, at comparably little computational cost. We observe appreciable differences between methods and across datasets and that no approach dominates all others. In summary, scds presents a scalable, competitive approach that allows for doublet annotation of datasets with thousands of cells in a matter of seconds.Availability and implementationscds is implemented as a Bioconductor R package (doi: 10.18129/B9.bioc.scds).Supplementary informationSupplementary data are available at Bioinformatics online.
| S-EPMC7703774 | biostudies-literature
Project description:Background: Single cell RNA sequencing (scRNA-seq) has rapidly gained popularity for profiling transcriptomes of hundreds to thousands of single cells. This technology has led to the discovery of novel cell types and revealed insights into the development of complex tissues. However, many technical challenges need to be overcome during data generation. Due to minute amounts of starting material, samples undergo extensive amplification, increasing technical variability. A solution for mitigating amplification biases is to include unique molecular identifiers (UMIs), which tag individual molecules. Transcript abundances are then estimated from the number of unique UMIs aligning to a specific gene, with PCR duplicates resulting in copies of the UMI not included in expression estimates. Methods: Here we investigate the effect of gene length bias in scRNA-Seq across a variety of datasets that differ in terms of capture technology, library preparation, cell types and species. Results: We find that scRNA-seq datasets that have been sequenced using a full-length transcript protocol exhibit gene length bias akin to bulk RNA-seq data. Specifically, shorter genes tend to have lower counts and a higher rate of dropout. In contrast, protocols that include UMIs do not exhibit gene length bias, with a mostly uniform rate of dropout across genes of varying length. Across four different scRNA-Seq datasets profiling mouse embryonic stem cells (mESCs), we found the subset of genes that are only detected in the UMI datasets tended to be shorter, while the subset of genes detected only in the full-length datasets tended to be longer. Conclusions: We find that the choice of scRNA-seq protocol influences the detection rate of genes, and that full-length datasets exhibit gene-length bias. In addition, despite clear differences between UMI and full-length transcript data, we illustrate that full-length and UMI data can be combined to reveal the underlying biology influencing expression of mESCs.
| S-EPMC5428526 | biostudies-literature
Project description:BACKGROUND: In addition to genome sequencing, accurate functional annotation of genomes is required in order to carry out comparative and evolutionary analyses between species. Among primates, the human genome is the most extensively annotated. Human miRNA gene annotation is based on multiple lines of evidence including evidence for expression as well as prediction of the characteristic hairpin structure. In contrast, most miRNA genes in non-human primates are annotated based on homology without any expression evidence. We have sequenced small-RNA libraries from chimpanzee, gorilla, orangutan and rhesus macaque from multiple individuals and tissues. Using patterns of miRNA expression in conjunction with a model of miRNA biogenesis we used these high-throughput sequencing data to identify novel miRNAs in non-human primates. RESULTS: We predicted 47 new miRNAs in chimpanzee, 240 in gorilla, 55 in orangutan and 47 in rhesus macaque. The algorithm we used was able to predict 64% of the previously known miRNAs in chimpanzee, 94% in gorilla, 61% in orangutan and 71% in rhesus macaque. We therefore added evidence for expression in between one and five tissues to miRNAs that were previously annotated based only on homology to human miRNAs. We increased from 60 to 175 the number miRNAs that are located in orthologous regions in humans and the four non-human primate species studied here. CONCLUSIONS: In this study we provide expression evidence for homology-based annotated miRNAs and predict de novo miRNAs in four non-human primate species. We increased the number of annotated miRNA genes and provided evidence for their expression in four non-human primates. Similar approaches using different individuals and tissues would improve annotation in non-human primates and allow for further comparative studies in the future.
| S-EPMC3328248 | biostudies-literature