Bridging the gap between reference and real transcriptomes

Morillon, Antonin; Gautheret, Daniel

doi:10.1186/s13059-019-1710-7

Cited by 51 publications

(39 citation statements)

References 90 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The human reference gene set of NCBI RefSeq contains 5.6 transcripts per coding locus, many more than the plant (1.7 t/l) or pig (3.0 t/l) reference sets. Yet the 113,000 transcripts of this reference set is half that found by RNA-seq assembly projects (Morillon & Gautheret 2019), with evidence of biological effects in those additional transcripts. With large effort, the human gene set is now the most accurate and complete of complex eukaryotes, implying that alternate and paralog transcripts are under-reported in most animal and plant gene sets.…”

Section: D: Value Of Accuracy For Alternates and Paralogsmentioning

confidence: 89%

“…Dougherty et al 2018 for cross-spliced paralogs). Long-read, single molecule sequencing methods have demonstrated value at recovering accurate transcripts, but as Morillon & Gautheret (2019) indicate, cost-effective and accurate short-read RNA data is an abundant resource for improved computational reconstruction of transcripts. The uncertainty in measurement accuracy of transcripts can be reduced with additional evidence evaluations, including replication over methods, and over biological samples.…”

Section: D: Value Of Accuracy For Alternates and Paralogsmentioning

confidence: 99%

See 1 more Smart Citation

Longest protein, longest transcript or most expression, for accurate gene reconstruction of transcriptomes?

Gilbert

2019

Preprint

View full text Add to dashboard Cite

Methods of transcript assembly and reduction filters are compared for recovery of reference gene sets of human, pig and plant, including longest coding sequence with EvidentialGene, longest transcript with CD-HIT, and most RNA-seq with TransRate. EvidentialGene methods are the most accurate in recovering reference genes, and maintain accuracy for alternate transcripts and paralogs. In comparison, filtering large over-assemblies by longest RNA measures, and most RNA-seq expression measures, discards a large portion of accurate models, especially alternates and paralogs. Accuracy of protein calculations is compared, with errors found in popular methods, as is accuracy of transcript assemblers. Gene reconstruction accuracy depends upon the underlying measurements, where protein criteria, including homology among species, have the strength of evolutionary biology that other criteria lack. EvidentialGene provides a gene reconstruction algorithm that is consistent with genome biology.Accurate gene set reconstruction 2019 October p. 1Some results of this comparison are obvious : longest transcript filter has longer transcripts, longest CDS filter recovers longer proteins, and most RNA-seq filter yields greater expression measures, compared to the others. The underlying question is which approach returns the most accurate gene information, consistent with efficient reduction to levels at which external evidence can be applied? Where results of these reduction filters differ, the one with greater biological information and phylogenetic validation, is presumed to be of more interest and utility to biologists.Proteins are evolutionarily conserved, functionally understandable biological information. The biological meaning of coding genes is in their coding sequence, so that discrepancies in CDS versus RNA quality measures favor the CDS measure. RNA-seq expression measures have technical imprecisions, with less direct biological meaning when these qualities deviate from coding sequence quality. The corre-Accurate gene set reconstruction 2019 October p. 2 spondence of protein-related quality measures, including protein size and homology, to biological protein recovery, via proteomics experiments, is known to be well above the correspondence with of expression quality measures (Tress et al. 2017).This report details use of these three filters to select accurate and complete gene sets from supersets of gene models that contain many accurate genes, plus redundant and less accurate models. Important as well, accurate coding sequence translation is discussed, and the value of several self-referential quality measures for accurate gene set reconstruction. Not considered here are chromosomal evidence, details of homology and external evidence, nor methods of non-coding gene validation. Those are important for accurate gene set reconstruction, and can be applied to the limited-palette results of self-referential draft gene sets. Self-referential gene set reconstruction, when done properly, is an efficient, data-intensive, first step i...

show abstract

Section: D: Value Of Accuracy For Alternates and Paralogsmentioning

confidence: 89%

Section: D: Value Of Accuracy For Alternates and Paralogsmentioning

confidence: 99%

Longest protein, longest transcript or most expression, for accurate gene reconstruction of transcriptomes?

Gilbert

2019

Preprint

View full text Add to dashboard Cite

show abstract

“…These mechanisms occur in various physiological or pathological processes such as development, cancer and immunity [4,5,6]. To date, there is no finite list of long non-coding isoforms, making it difficult to construct a complete lncRNA catalogue due to the high number of transcripts and their tissue-specific expression [7,8]. The absence of a complete catalogue makes it difficult to establish a comprehensive lncRNA expression profile.Thus, currently, the best strategy for the study of lncRNAs consists in the prediction of transcripts from a selection of RNAseq data in a tissue-specific condition.…”

Section: Introductionmentioning

confidence: 99%

Detailed analysis of public RNAseq data and long non-coding RNA: a proposed enhancement to mesenchymal stem cell characterisation

Riquier

Mathieu

Boureux

et al. 2020

Preprint

View full text Add to dashboard Cite

The development of RNA sequencing (RNAseq) and corresponding emergence of public datasets have created new avenues of transcriptional marker search. The long non-coding RNAs (lncRNAs) constitute an emerging class of transcripts with a potential for high tissue specificity and function. Using a dedicated bioinformatics pipeline, we propose to construct a cell-specific catalogue of unannotated lncRNAs and to identify the strongest cell markers. This pipeline uses ab initio transcript identification, pseudoalignment and new methodologies such as a specific k-mer approach for naive quantification of expression in numerous RNAseq data.For an application model, we focused on Mesenchymal Stem Cells (MSCs), a type of adult multipotent stem-cells of diverse tissue origins. Frequently used in clinics, these cells lack extensive characterisation. Our pipeline was able to highlight different lncRNAs with high specificity for MSCs. In silico methodologies for functional prediction demonstrated that each candidate represents one specific state of MSCs biology. Together, these results suggest an approach that can be employed to harness lncRNA as cell marker, showing different candidates as potential actors in MSCs biology, while suggesting promising directions for future experimental investigations.

show abstract

“…Consequently, it is important to understand how TEs contribute to the coding and non-coding transcriptome (You et al 2017;Ma et al 2019;Morillon and Gautheret 2019), if hPSCs are to be used in cell replacement therapy (Fort et al 2014;Schumann et al 2019). However, due to limitations in short-read based sequencing, accurate transcriptome maps have been elusive.…”

Section: Introductionmentioning

confidence: 99%

Transposable Element-Gene Splicing Modulates the Transcriptional Landscape of Human Pluripotent Stem Cells

Babarinde

et al. 2020

Preprint

View full text Add to dashboard Cite

Transposable elements (TEs) occupy nearly 50% of mammalian genomes and are both potential dangers to genome stability and functional genetic elements. TEs can be expressed and exonised as part of a transcript, however, their full contribution to the transcript splicing remains unresolved. Here, guided by long and short read sequencing of RNAs, we show that 26% of coding and 65% of non-coding transcripts of human pluripotent stem cells (hPSCs) contain TEs. Different TE families have unique integration patterns with diverse consequences on RNA expression and function. We identify hPSC-specific splicing of endogenous retroviruses (ERVs) as well as LINE L1 elements into protein coding genes that generate TE-derived peptides. Finally, single cell RNA-seq reveals that proliferating hPSCs are dominated by ERV-containing transcripts, and subpopulations express SINE or LINE-containing transcripts. Overall, we demonstrate that TE splicing modulates the pluripotency transcriptome by enhancing and impairing transcript expression and generating novel transcripts and peptides.

show abstract

Bridging the gap between reference and real transcriptomes

Cited by 51 publications

References 90 publications

Longest protein, longest transcript or most expression, for accurate gene reconstruction of transcriptomes?

Longest protein, longest transcript or most expression, for accurate gene reconstruction of transcriptomes?

Detailed analysis of public RNAseq data and long non-coding RNA: a proposed enhancement to mesenchymal stem cell characterisation

Transposable Element-Gene Splicing Modulates the Transcriptional Landscape of Human Pluripotent Stem Cells

Contact Info

Product

Resources

About