Gene expression or gene regulation studies often assume one gene expresses one mRNA. However, contrary to the conventional idea, a single gene in mammalian cells can express multiple transcript variants translated into several different proteins. The transcript variants are generated through transcription from alternative start sites and alternative post-transcriptional processing of the precursor mRNA (pre-mRNA). In addition, gene mutations and RNA editing further enhance the diversity of the transcript variants. The transcript variants can encode proteins with various domains, expanding the functional repertoire of a single gene. Some transcript variants may not encode proteins but function as non-coding RNAs and regulate gene expression. The expression level of the transcript variants may vary between cell types or within the same cells under different biological conditions. Transcript variants are characteristic of cell differentiation in a particular tissue, and the variants may play a key role in normal development and aging. Studies also reported that some transcript variants may have roles in disease pathogenesis. The biological significances urge studying the complexity of gene expression at the transcript level. This article updates the molecular basis of transcript variants in mammalian cells, including the formation mechanisms and potential roles in host biology. Gaining insight into the transcript variants will not only identify novel mechanisms of gene regulation but also unravel the role of the variants in health and disease.

RNA sequencing (RNA-Seq) has become a widely adopted technique for studying gene expression. However, conventional RNA-Seq analyses rely on gene expression (GE) values that aggregate all the transcripts produced under a single gene identifier, overlooking the complexity of transcript variants arising from different transcription start sites or alternative splicing. Transcript variants may encode proteins with diverse functional domains, or noncoding RNAs. This study explored the implications of neglecting transcript variants in RNA-Seq analyses. Among the 1334 transcription factor (TF) genes expressed in mouse embryonic stem (ES) or trophoblast stem (TS) cells, 652 were differentially expressed in TS cells based on GE values (365 upregulated and 287 downregulated, ≥absolute 2-fold changes, false discovery rate (FDR) p-value ≤ 0.05). The 365 upregulated genes expressed 883 transcript variants. Further transcript expression (TE) based analyses identified only 174 (<20%) of the 883 transcripts to be upregulated. The remaining 709 transcripts were either downregulated or showed no significant changes. Meanwhile, the 287 downregulated genes expressed 856 transcript variants and only 153 (<20%) of the 856 transcripts were downregulated. The other 703 transcripts were either upregulated or showed no significant change. Additionally, the 682 insignificant TF genes (GE values < absolute 2-fold changes and/or FDR p-values > 0.05) between ES and TS cells expressed 2215 transcript variants. These included 477 (>21%) differentially expressed transcripts (276 upregulated and 201 downregulated, ≥absolute 2-fold changes, FDR p-value ≤ 0.05). Hence, GE based RNA-Seq analyses do not represent accurate expression levels due to divergent transcripts expression from the same gene. Our findings show that by including transcript variants in RNA-Seq analyses, we can generate a precise understanding of a gene’s functional and regulatory landscape; ignoring the variants may result in an erroneous interpretation.