Recycling and metabolic flexibility dictate life in the lower oceanic crust

Contesting the evidence for gene expression in lower oceanic crust

“…Li et al . 1 reported gene expression from a microbial biosphere in the lower ocean crust, from samples that have cell concentrations as low as 100 cells per cm −3 . This would be incredible, since that would be the lowest biomass sample ever analyzed for gene expression for a subsurface sample by several orders of magnitude.…”

“…The authors discuss Pseudomonas as being one of the most active groups in their samples 1 , which is concerning because Pseudomonas is one of the most common contaminants in molecular reagent kits 2 . This prompted me to check the annotations of the entire dataset provided by the STAT annotation pipeline using the NCBI RefSeq database available on the short read archive (SRA) website.…”

“…The high amount of seawater contamination is exacerbated by the mRNA amplification method used by Li et al 1 . The Ovation RNA-Seq kit that the authors used has a high fidelity SPIA polymerase for amplifying nucleic acids, and is highly sensitive to DNA contamination when used on minute amounts of RNA 10 .…”

“…Li et al 1 write that they took extensive measures to reduce contamination during sampling, and performed negative controls in the form of extraction blanks of DNA and RNA. They write in the Methods that they removed sequences from “…taxa known to be common molecular kit reagent contaminants 46,56 ”, which in the cited Salter et al 2 paper includes Deinococcus-Thermus and Pseudomonas .…”

“…Despite being included in the database (Figure 1B), sequences annotated to MAGs and SAGs from inhabitants of crustal subsurface habitats such as ammonia oxidixing Thaumarchaeota, Bathyarchaeota, sulfide oxidizing Epsilonproteobacteria, and iron oxidizing Zetaproteobacteria 4,5,12 are notably missing in the gene expression dataset of Li et al 1 (Figure 1B). While all subsurface gene expression data are prone to contamination due to the low biomass as an unavoidable challenge, it is nevertheless expected in the case of truly active subsurface microbial communities that at least some sequences derived the typical inhabitants of these extreme environments should be represented with reliable levels of sequencing coverage.…”