Arbuscular mycorrhizal fungi (AMF) are important symbionts of plants. Recently, studies of the AMF Rhizophagus irregularis recorded within-isolate genetic variation that does not completely match the proposed homokaryon or heterokaryon state (where heterokaryons comprise a population of two distinct nucleus genotypes). We re-analysed published data showing that bi-allelic sites (and their frequencies), detected in proposed homo- and heterokaryote R. irregularis isolates, were similar across independent studies using different techniques. This indicated that observed within-fungus genetic variation was not an artefact of sequencing and that such within- fungus genetic variation possibly exists. We then looked to see if bi-allelic transcripts from three R. irregularis isolates matched those observed in the genome as this would give a strong indication of whether bi-allelic sites recorded in the genome were reliable variants. In putative homokaryon isolates, very few bi-allelic transcripts matched those in the genome. In a putative heterokaryon, a large number of bi-allelic transcripts matched those in the genome. Bi-allelic transcripts also occurred in the same frequency in the putative heterokaryon as predicted from allele frequency in the genome. Our results indicate that while within-fungus genome variation in putative homokaryon and heterokaryon AMF was highly similar in 2 independent studies, there was little support that this variation is transcribed in homokaryons. In contrast, within-fungus variation thought to be segregated among two nucleus genotypes in a heterokaryon isolate was indeed transcribed in a way that is proportional to that seen in the genome.