Correction to: Nature https://doi.org/10.1038/s41586-020-2075-5 Published online 11 March 2020

It has been brought to the authors’ attention that the pipeline steps for metatranscriptome contamination removal could miss mapping any sample reads to control reads that didn’t assemble into the control co-assembly. Therefore, an additional analysis of the manually checked transcripts behind the themes discussed in the paper was performed using Magic-BLAST, a mapping tool that allows detection and removal of reads in sample data sets that are mapped to by control reads. Transcripts discussed in the paper flagged by Magic-BLAST were removed from the dataset if they recruited one or more reads from a control sample at 98% local identity over 50% or more of the alignment. The new determinations did not compromise any of the pathways discussed in our paper, but removal of reads (~19%) changed the calculated FPKM values reported throughout Supplementary Table 4 of the original paper. Although we advise against using FPKM values to compare expression levels of selected genes between samples for amplified data sets, this amendment provides updated FPKM values for Supplementary Table 4, and highlights a few instances where all transcripts for individual genes within categories were removed. At the same time, we updated taxonomic annotations as of spring 2020, as databases are constantly improving, and where possible, we now provide synonyms and abbreviations for the reported functions in an effort to make the table more useful to the reader. Additional useful detail is also provided about the analysis of metatranscriptome data within the selected categories of metabolism discussed in the paper, and about the clean room facility used for the cell counts.

As Supplementary Information files to this amendment please find (1), a brief Supplementary Methods describing the clean room cell counting facility used for cell counts, providing detection limits of the instrument used for the ATP assays, and explaining details of the additional analysis of the mRNA data, and (2), a Supplementary Discussion file showing minor text updates to the mRNA section of our original Supplementary Discussion reflecting the revised Supplementary Table 4 (also included in this amendment). Fasta sequences for contigs reported in Supplementary Table 4 within different categories of metabolism are provided as a Supplementary Data file. A note has been added to the dataset names for our originally deposited transcriptome data sets under BioProject PRJNA497074 to make it clear that they are raw data, and contaminating sequences have not been removed. Sample data set names now include “…control sample sequences not removed”.