Critical evaluation of the FANTOM3 non-coding RNA transcripts

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• The FANTOM3 project identified numerous non-coding RNA (ncRNA) transcripts.
• Understanding the nature and origin of these ncRNAs is crucial for deciphering their biological roles.
• Distinguishing genuine ncRNAs from experimental artefacts is a significant challenge.

Purpose of the Study:

• To analyze the genomic distribution of FANTOM3 ncRNAs relative to protein-coding genes.
• To identify and filter a subset of high-quality ncRNA transcripts suitable for further experimental investigation.
• To characterize potential artefacts such as internal priming and overlap with protein-coding regions.

Main Methods:

• Bioinformatic analysis of 38,129 putative ncRNAs from the RIKEN dataset.
• Classification of transcripts based on their genomic location (sense, antisense, intronic, intergenic).
• Filtering criteria included non-overlap with protein-coding genes, absence of long open reading frames (ORFs), and exclusion of internally primed transcripts.

Main Results:

• The initial dataset comprised 41% sense, 6% antisense, 24% intronic, and 29% intergenic transcripts.
• A significant fraction (47%) of transcripts showed signs of potential internal priming.
• A filtered subset of 8535 transcripts was generated, free from common artefacts and overlapping protein-coding genes.
• This filtered set includes 53% of known ncRNAs from RNAdb and 6523 novel transcripts.

Conclusions:

• Bioinformatic filtering effectively removed potential artefacts from the FANTOM3 non-coding dataset.
• The refined dataset of 8535 transcripts represents a high-confidence set for investigating novel non-coding RNA functions.
• This work provides a valuable resource for experimental studies using hybridization-based techniques.