PURPOSE: Guidelines by professional organizations for assessing variant pathogenicity include the recommendation to utilize biologically relevant transcripts; however, there is variability in transcript selection by laboratories. METHODS: We describe three patients whose genomic results were incorrect, because alternative transcripts and tissue expression patterns were not considered by the commercial laboratories. RESULTS: In individual 1, a pathogenic coding variant in a brain-expressed isoform of CKDL5 was missed twice on sequencing, because the variant was intronic in the transcripts considered in analysis. In individual 2, a microdeletion affecting KMT2C was not reported on microarray, since deletions of proximal exons in this gene are seen in healthy individuals; however, this individual had a more distal deletion involving the brain-expressed KMT2C isoform, giving her a diagnosis of Kleefstra syndrome. Individual 3 was reported to have a pathogenic variant in exon 10 of OFD1 on exome, but had no typical features of the OFD1-related disorders. Since exon 10 is spliced from the more biologically relevant transcripts of OFD1, it was determined that he did not have an OFD1 disorder. CONCLUSION: These examples illustrate the importance of considering alternative transcripts as a potential confounder when genetic results are negative or discordant with the phenotype.