Week 3 - Antarctic samples are in!

This was a very short, yet productive week in the lab. The Mt. Erebus/Warren cave metagenome sequences finally arrived from Illumina sequencing Friday evening and I began my work Monday querying the samples for coxL, the gene encoding the enzyme carbon monoxide dehydrogenase (CODH), then assembling those genes for analysis.

The metagenome first needed to be concatenated (linked end-to-end), because it was sequenced in small segments. I learned a quick unix command to accomplish this. In the process of Illumina sequencing the ends of the sequences often deteriorate in quality, which can make assembly problematic. I used the Sickle tool (https://github.com/najoshi/sickle) to trim the degenerate 3' and 5'-ends of the concatenated sequences. Sickle uses a "sliding window" to walk the length of the single end reads and determine when the quality is low enough to trim and discard the ends. After trimming, I used the IMG coxL database I created in week 1 to query the metagenome for coxL genes and extracted those sequences for assembly. After waiting 4 hours for the 120,549 reads to assemble in Geneious (v. 5.6), 3528 contigs were generated! I analyzed the contigs and sorted them by the greatest number of reads per contig hoping to see great consensus coverage over the metagenome (a good indicator that coxL is sufficiently present in the community). However, to my surprise the contigs were gap-filled and consensus coverage was poor. Something definitely looked wrong…

Rick came in Thursday with a solution. Apparently, the transposon that was used to indiscriminately cut the metagenomic DNA into smaller fragments for sequencing was present in many of the reads. After identifying the transposon nucleotide sequence (CTGTCTCTTATA), I used BioEdit (v. 7.1.3) to search for reads containing the transposon. Once these reads were identified, I removed the transposon and all of the bases following it. I then reassembled the reads in Geneious and generated contigs with very high overlap and great consensus coverage. I extracted the consensus sequences from the highest quality contigs and ran blastx searches on the National Center for Biotechnology Information (http://blast.ncbi.nlm.nih.gov/Blast.cgi) database of annotated microbial genomes. The search resulted in hits only to known carbon monoxide dehydrogenase, large subunit genes from both cultured and uncultured species. Friday I reassembled the reads with more robust parameters and developed a protocol for determining which contigs I would use in my analysis. I decided to allow the contigs to have a minimum of 25 reads in any portion of the sequence to generate higher quality consensus sequences. Also, I constrained my analysis to contigs with a minimum of 1000 reads to ensure each contig I analyzed was representative of the community. With these parameters I was restricted to using 4 high quality contigs out of the 2337 that were generated upon assembly.

The coxL gene appears to be present in significant amounts in the metagenome and thus carbon monoxide oxidation might prove to be a significant contributor to the energy and carbon requirements of the microbial communities in Warren cave. This would be an interesting insight into the metabolic potential of these microbes as the capacity for carbon monoxide oxidation has never been discovered in these types of environments before. Microbial carbon monoxide oxidation has been shown to be diverse in its distribution (Cunliffe 2011; Dunfield and King 2004; Weber and King 2012), yet the full potential of carbon monoxide oxidation in nature is severely understudied. Much of the work having been done on culture-able bacteria, is limited in scope and is likely not representative of carbon monoxide oxidation in vivo (King and Weber 2007). My project may provide insights into how diverse a process CO-oxidation is and expand the knowledge of CO-oxidizing bacteria not previously known to use CO metabolically.

Cunliffe M (2011) Correlating carbon monoxide oxidation with cox genes in the abundant Marine Roseobacter Clade. The ISME Journal 5: 685-691
Dunfield KE, King GM (2004) Molecular analysis of carbon monoxide-oxidizing bacteria associated with recent Hawaiian volcanic deposits. Applied and Environmental Microbiology 70: 4242-4248
King GM, Carolyn WF (2007) Distribution, diversity and ecology of aerobic CO-oxidizing bacteria. Nature Reviews Microbiology 5: 107-18
Weber CF, GM King (2012) The phylogenetic distribution and ecological role of carbon monoxide oxidation in the genus Burkholderia. FEMS Microbial Ecology 79: 167-175