N 1 -methyladenosine (m 1 A) was recently identified as a new mRNA modification based on its mapping to the 5' UTRs of thousands of mRNAs with an m 1 A-binding antibody. More recent studies have confirmed the prevalence of m 1 A, while others have questioned it. To address this discrepancy, we mapped m 1 A using ultra-deep RNA-Seq datasets based on m 1 A-induced misincorporations during reverse transcription. Using this approach, we find m 1 A only in the mitochondrial MT-ND5 transcript. In contrast, when we mapped m 1 A antibody-binding sites at single-nucleotide resolution, we found binding to transcription start nucleotides in mRNA 5' UTRs. Using different biochemical assays, we find that m 1 A is not present at these sites. Instead, we find that the m 1 A antibody exhibits m 1 A-independent binding to mRNA cap structures. We also tested a new and independently derived m 1 A antibody. We show that this m 1 A antibody lacks m 7 G cap-binding cross-reactivity, and notably does not map to 5' UTRs in the transcriptome. Our data demonstrate that high-stoichiometry m 1 A sites are rare in the transcriptome and that previous mapping of m 1 A to mRNA 5' UTRs are due to unintended binding of the m 1 A antibody to m 7 G cap structure in mRNA.Given the inconsistency in the different antibody-dependent m 1 A-mapping methods (Extended Data Fig. 1a-c), we sought to use an antibody-independent approach to detect m 1 A at singlenucleotide resolution in mRNA. For this approach, we took advantage of ultra-deep RNA-seq datasets and the fact that m 1 A is a "hard stop" modification, meaning it typically arrests cDNA synthesized by standard reverse transcriptases 7,8 (Extended Data Fig. 1d). However, SuperScript III will read through m 1 A at low frequency, resulting in misincorporations that are variable and sequence dependent 7,8 . Most m 1 A-induced misincorporations are A!T transitions that can be detected by sequencing the cDNA 7,8 . This approach can also detect other hard stop modifications, such as 3-methylcytidine (m 3 C), 3-methyluridine (m 3 U), N 2 ,N 2 -dimethylguanosine, and 1-methylguanosine, since these also produce misincorporations 7 . Therefore, the presence of misincorporated nucleotides can directly localize the m 1 A and other hard stop nucleotides in sequencing data 7,8 .