Hidden gene fusions reshape the tuberculosis genome

Researchers have long considered the genome of Mycobacterium tuberculosis (Mtb) to be unusually stable and clonal, with evolution driven mainly by single nucleotide polymorphisms (SNPs) and small insertions or deletions (INDELs). Yet clinical strains often behave very differently from laboratory strains, showing altered growth, virulence, and metabolism that are hard to explain only by SNPs and INDELs. Nikhil Bhalla and colleagues set out to test whether pieces of distant genes can be brought together inside the same RNA molecule—so-called fusion transcripts—by natural recombination within the Mtb genome. The Mtb genome contains transposases, redundant genes, repetitive DNA, integrases, and remnants of lysogenized mycobacteriophages, all of which could promote intragenomic recombination. Using public RNA-seq data from clinical Mtb isolates, the team optimized a straightforward fusion-calling strategy based on established bioinformatics methods to look for long-distance fusion events between non-operonic genes. Their goal was to detect robust evidence of fusion transcripts and to check whether such events might be lineage-specific or linked to particular mobile genetic elements.

The study tested three complementary computational approaches: split read alignment, repurposing STAR chimera, and transcript de novo assembly. The researchers evaluated intersections of fusion calls across these methods to find the most reliable signals, and they report that the overlap between the split reads approach and the repurposed STAR chimera had high performance (F1 > 0.9). Applying the optimized strategy to real RNA-seq datasets showed consistent sequence characteristics, clustering patterns, and gene burden for both operonic and long-distance gene fusions across two independent data sets, demonstrating robustness. The fusion transcripts displayed lineage specificity and pointed to indirect involvement of transposases and specific transposition accessory genes—Rv1199c, Rv2512c, Rv3115, Rv0395, Rv2808, and Rv3327—in intragenomic recombination that produces fusion transcripts. Most fusions were within transposases, PPE, and PE_PGRS family proteins, while some isolated fusions involved genes in the MoCo pathway, vesicle transport, and lipid turnover.

Taken together, these findings indicate that natural recombination within Mtb can create fusion transcripts that may produce fusion proteins, co-regulate proteins, or disrupt normal genes. This challenges the idea that clinical Mtb genomes are largely immutable and driven only by SNPs and INDELs. The presence of transposases and accessory recombination genes means distant parts of the genome can be rearranged without outside horizontal gene transfer, altering gene repertoires and copy numbers. The study’s optimized, easy-to-implement fusion-calling workflow provides a practical tool for other researchers to search RNA-seq data for such events, and the lineage-specific patterns suggest these fusions could contribute to the phenotypic diversity seen in clinical isolates. In short, genome plasticity mediated by intragenomic recombination may be an underappreciated force shaping Mtb biology in the clinic.