Written by Charlotte Harrison, Science Writer
Cancer transcriptomics has transformed our understanding of tumour biology. Yet even the most up-to-date transcriptomic studies using RNA sequencing (RNS-seq) mostly quantify the expression of annotated genes listed in a reference transcriptome. In short, current methods only interpret the better-understood data.
This means a wide range of non-standard RNA information is ignored – such as mRNA isoforms, non-coding RNAs, and endogenous retroelements. Could we be missing information that improves our understanding of cancer and helps to develop new therapies?
Now, a study published in NAR Cancer, which collected non-standard RNA information from tumour samples, has taken RNA exploration into new territory.
To go where (probably) no one has gone before
The researchers used a previously developed computational method to obtain information on non-standard RNA directly from raw RNA-seq data. They compared data from normal samples with two data sets of lung adenocarcinoma. In particular, their analysis focused on events that are ignored in studies using conventional transcriptomics, and assessed the value of these events as biomarkers, tumour classifiers and survival predictors.
Overall, their results revealed a collection of novel tumour-specific RNAs, including long intergenic non-coding RNAs, novel splice variants, intron retentions and viral- and bacterial-derived transcripts. Each of these RNAs contributed to a different, important aspect of tumour identity.
Discovering new worlds (of biomarkers and survival predictors)
A novel aspect of the study’s reference-free RNA-seq analysis was its capacity to identify novel cancer drivers or biomarkers. In essence, the approach casts a wider net for biomarker discovery, as it collects all events independently of their origin.
The authors identified a subset of RNA events — including those in intergenic locations and non-coding RNAs — that were not expressed in normal tissue. These events could be potential sources of tumour-specific antigens, and as such might one day be used as novel biomarkers or targets for treatment. The authors note that targeting these tumour-specific antigens would potentially benefit large patient groups.
Moreover, the study identified several sequences that could predict overall survival. These were mostly enriched in repeats, especially HERV elements (sequences obtained from exogenous retroviral infections throughout the evolution of the human genome) and transposable L1 elements.
Using this method of reference-free transcriptomics could be particularly valuable in cancer, as each individual tumour has a unique transcriptome that differs in unpredictable ways from that of normal tissue. Moreover, the depth of information collected by this method incentivizes its use to explore non-standard RNAs in other diseases.
Image Credit: Canva
