Understanding FEELnc Tool Results and lncRNA Functional Analysis

Question

[Mian Numan]

Hi there, I’m a molecular biologist and I’m a bit confused about something we did. We used the FEELnc tool to predict lncRNA, and in the output files, we see the lncRNA MSTRG along with some genes named transcript: AT2G29686.3. What does this mean? In lncRNA functional analysis (GO), we look for genes that are expressed similarly to the lncRNA we’re considering. If we find co-expressed genes, we perform a Gene Ontology analysis. But I’m not sure if that’s what we’re doing here. Can you help me understand?

Answer 1

[Kevin Blighe]

Apologies that nobody has yet responded to you.

In FEELnc output (built on StringTie-assembled transcripts), "MSTRG" (e.g., MSTRG.x.y) denotes novel, de novo assembled transcripts predicted as lncRNAs—StringTie-generated candidates that FEELnc classifies as non-coding based on length, coding potential, and genomic context. These are often intergenic or genic relative to known genes.

"transcript: AT2G29686.3" is a standard ID for a known Arabidopsis protein-coding gene transcript (TAIR locus). In FEELnc's classifier output, it likely indicates an interaction: your lncRNA (MSTRG) overlaps, is antisense to, or is proximal to this gene, suggesting potential cis-regulation.

For functional analysis: Yes, co-expression (e.g., via Pearson correlation on RNA-seq data) identifies similarly expressed protein-coding genes, followed by GO enrichment on those genes to infer lncRNA roles (guilt-by-association). This is standard for novel lncRNAs like yours. Check FEELnc's .interactions.tsv for details.

Kevin