Today's paper is: 

Cissé, O. H., Pagni, M., & Hauser, P. M. (2012). De novo assembly of the Pneumocystis jirovecii genome from a single bronchoalveolar lavage fluid specimen from a patient. mBio, 4(1), e00428-12. American Society for Microbiology. doi:10.1128/mBio.00428-12.

Pneumocystis jirovecii is a fungus that isn't usually pathogenic but is deadly to immunocompromised patients, most prominently those with AIDS.  As with many other parasitic microbes colonizing human tissues, P. jirovecii can't be cultured in the lab. This disadvantage makes working with P. jirovecii nearly impossible if researchers seek to use genetic or molecular techniques. This looks like a case for sequencing! Wait, no, hold on. Researchers generally need to culture microbes to get enough genomic DNA for sequencing. The authors of this paper bypass that issue by collecting the fungus direct from the lungs of infected patients (they also used immunoprecipitation and random amplification, so this by no means an instance of single-cell sequencing).

Collecting samples of bronchoalveolar lavage fluid (or, if you prefer, the ridiculous-sounding acronym BALFs) provides plenty of fungus to work with. The samples are also rich in other genetic material from human cells, bacteria, viruses, and on occasion other species of Pneumocystis. The authors had to extract the P. jirovecii genome from the mess -- luckily, the genome of related species and rat pathogen P. carinii has already been sequenced and partially assembled. Process of elimination allowed for removal of the non-Pneumocystis sequences, then genome comparison and RNAseq helped finish and annotate the rest. More than just a de novo genome assembly using a messy metagenomic data source, this work is an example of how availability of massive quantities of genomic data make some projects entirely feasible.