Short talk - Deciphering the “m6A Code” via Antibody-Independent Quantitative Profiling

Slides

1: Intro

My name is Miguel Garcia. My research centers on m6A, the most abundant mRNA modification on eukaryotes; implicated in development, physiology, and disease. A major limitation in the field has been the inability to quantify its stoichiometry at individual sites in a high-throughput manner. Blinding us to the fine-grained resolution in which m6A may operate.

2: MAZTER-seq methodology

To alleviate this limitation, we developed MAZTER-seq, coupling RNA-seq with the ability of the MazF RNAse to cleave at the unmethylated ACA motif, but not at its m6A-methylated counterpart. For analysis of MAZTER-seq data, we developed a computational pipeline called MAZTER-mine, which processes raw data and calculates the mazF “Cleavage efficiency” for each ACA site in the transcriptome. Using in-vitro transcripts harboring a single m6A site, we generated spike-ins with different stoichiometries of m6A, for which MAZTER-seq yielded highly correlated measurements. Next, an example of a known m6A site in a wild-type yeast strain, in which cleavage by mazF is substantially reduced compared to a double knock-out strain of the m6A methyltransferase IME4.

3. m6A stoichiometry is predictable across the transcriptome

After establishing MAZTER-seq, we compared all the ACA sites of the IME4 knock-out yeast to its control. Doing this, we detected significant differences for 410 sites, including 56 previously known sites. Using both known and newly detected m6A sites, we used their sequence, along with additional data, to fit a linear regression. Our results show that measured and predicted stoichiometries have a high correlation, explaining almost fifty percent of the variation in stoichiometry levels. Moreover, we performed an analog experiment in mouse embryonic stem cells, achieving comparable results when knocking out the methyltransferase METTL3. Graphical depictions of the models resonate with each other, as both models share similar coefficients to calculate m6A stoichiometries.

Epilogue

My results, thus show that: MAZTER-seq allows systematic m 6 A quantification at single-nucleotide resolution in diverse biological settings. And, m 6 A stoichiometry is “hard-coded” by a simple, predictable, and conserved code.