With advances in DNA sequencing and miniaturized molecular biology workflows, rapid and affordable sequencing of single-cell genomes has become a reality. Compared to 16S rRNA gene surveys and shotgun metagenomics, large-scale application of single-cell genomics to whole microbial communities provides an integrated snapshot of community composition and function, directly links mobile elements to their hosts, and enables analysis of population heterogeneity of the dominant community members. To that end, we sequenced nearly 500 single-cell genomes from a low diversity hot spring sediment sample from Dewar Creek, British Columbia, and compared this approach to 16S rRNA gene amplicon and shotgun metagenomics applied to the same sample. We found that the broad taxonomic profiles were similar across the three sequencing approaches, though several lineages were missing from the 16S rRNA gene amplicon dataset, likely the result of primer mismatches. At the functional level, we detected a large array of mobile genetic elements present in the single-cell genomes but absent from the corresponding same species metagenome-assembled genomes. Moreover, we performed a single-cell population genomic analysis of the three most abundant community members, revealing differences in population structure based on mutation and recombination profiles. While the average pairwise nucleotide identities were similar across the dominant species-level lineages, we observed differences in the extent of recombination between these dominant populations. Most intriguingly, the creek’s Hydrogenobacter sp. population appeared to be so recombinogenic that it more closely resembled a sexual species than a clonally evolving microbe. Together, this work demonstrates that a randomized single-cell approach can be useful for the exploration of previously uncultivated microbes from community composition to population structure.