Skip to main content
NCBI home page
ISME Communications logoLink to ISME Communications
. 2025 Sep 3;5(1):ycaf152. doi: 10.1093/ismeco/ycaf152

Genomic recovery from rare terrestrial microbes enabled by DNA-based GC-fractionation

Paul O Sheridan 1,2, Yiyu Meng 3, Dylan Bodington 4, David Coutts 5, Tom A Williams 6, Cécile Gubry-Rangin 7,✉
PMCID: PMC12477608  PMID: 41030375

Abstract

Genome reconstruction from metagenomic samples has dramatically increased our understanding of uncultivated lineages of life. However, untargeted metagenomic sequencing is biased towards the more abundant microbes, neglecting less abundant lineages playing important ecological roles, such as the ammonia-oxidising archaea. Here, we demonstrate that separating soil molecular DNA using a bisbenzimide-CsCl guanine-cytosine (GC)-content-based DNA fractionation approach separates microbial DNA along a GC-content gradient. The fractions from both extremes of the GC-content gradient possess different 16S rRNA gene composition than the original unfractionated DNA. The high diversity in the lower GC-content fractions (< 45%) contrasts with the higher DNA abundance in the higher GC-content fractions (50%–70%), highlighting the low GC fractions as an enriched source of rare microbe DNA. Metagenomic sequencing of specific low- and high-GC fractions enabled the reconstruction of 204 taxonomically diverse metagenome-assembled genomes from 31 microbial phyla, with at least 63 of these originating from rare (< 0.1% relative abundance) or very rare (< 0.01% relative abundance) microbial families. Therefore, this approach facilitates genomic assembly of rare taxa in resulting pseudo-communities. Ultimately, this technique enables a semi-targeted metagenomic approach to recover genomes from low-abundance microbes with GC-contents that differ significantly from the environmental microbial community of interest. As mounting evidence suggests that rare microbes drive critical ecosystem functions, this approach will facilitate a deeper understanding of their metabolic potential in the environment.

Keywords: GC-content, bisbenzimide, genome, MAG, 16S rRNA gene, rare, low abundance, bacteria, archaea, AOA

Graphical Abstract

Graphical Abstract.

Graphical Abstract

Open in a new tab

Introduction

The ability to reconstruct genome sequences from shotgun metagenomic reads has greatly increased the genomic representation of the tree of life [1, 2], enabling metabolic prediction of many uncultivated microbes. However, complex microbial communities comprise hundreds to thousands of taxonomic phylotypes at vastly uneven abundances, with a high number of rare species [3]. In typical metagenomic experiments, most reads will be derived from abundant microorganisms [4–6], with rare microbes being poorly represented in the resulting genomic datasets. This is a problem because low-abundance taxa may perform a critical role in ecosystem function, health, and stability [3]. For example, despite their low abundance, the ammonia-oxidising archaea (AOA) perform the first step of the nitrification pathway and, therefore, are key players in the nitrogen cycle in multiple environments, including soil, marine, freshwater, sediment, marine sponge, and geothermal ecosystems [7–9]. In addition, low-abundance microbes contribute to the temporal community dynamics of an ecosystem [3], as they represent a seedbank of diversity which can be reactivated from dormancy in response to environmental change [10]. Therefore, reconstructing genomes of low-abundance taxa is important to better understand the role and ecosystem function of those organisms, even enabling genomics-resolved cultivation [11].

Reconstructing medium- to high-quality [12] draft genomes (metagenomics-assembled genomes, MAGs) of low-abundance taxa can theoretically be achieved by extremely high sequencing depth and massive computational resources. However, DNA enrichment approaches offer several technical benefits, including enhanced sequencing sensitivity and reduced sequencing and assembly costs, which are aligned with a lower required sequencing depth. Multiple selective sequencing approaches have been previously developed to enrich the DNA of low-abundance taxa. For example, methylation-guided enrichment approaches selectively enrich the microbial DNA by removing the host DNA based on its higher methylation level [13]. In addition, targeted-sequencing approaches enrich the target DNA during the library preparation, including with CRISPR-Cas9 enrichment or hybrid capture [14]. These later approaches require prior knowledge of the targeted community, which is often difficult to obtain in many underexplored ecosystems. Recently, selective nanopore sequencing and associated bioinformatic scripts were used to achieve efficient enrichment of rare taxa in diverse ecosystems [15] paving the way for computational development to acquire genomes of low-abundance taxa. However, the development of empirical approaches to enrich the microbial DNA of low-abundance taxa without prior knowledge of the community’s genomic content would advance the field of rare microbiomes.

To fill this gap of knowledge, we present a wet-lab approach which enables DNA separation based on the guanine-cytosine (GC) content of the organismal DNA. We hypothesized that GC-content-based DNA fractionation reduces the microbial community complexity into restricted GC-content pseudo-communities before sequencing. Additionally, the non-intercalating DNA dye bisbenzimide (also known as Hoechst 33258) preferentially binds to A + T DNA regions, enabling greater separation of DNA fragments based on GC-content and determined buoyant density [16]. As most soil environmental DNA tends to have high GC-content [17], the proposed approach should enable the enrichment of DNA from rare, low GC-content genomes and the selective sequencing of low GC-content DNA organisms. Bisbenzimide-enhanced GC-content-based DNA fractionation has been demonstrated in diverse ecosystems including plants [18], chicken digesta [19], sediments, bioreactors [20] and cricket hindguts [21], and should therefore be applicable to any environment, especially those where the GC-content of the environmental DNA is not evenly distributed. To our knowledge, this is the first time this technique has been applied in genome-resolved metagenomics, and we demonstrated its suitability to reconstruct medium- and high-quality MAGs from rare (< 0.1% relative abundance) [22] and very rare (< 0.01% relative abundance) [23] microbial taxa in soil, a highly complex environment [24] (see Fig. 1).

Figure 1.

Figure 1

Open in a new tab

Conceptual representation of the bisbenzimide-CsCl DNA fractionation approach for genomic recovery of non-abundant microbes.

Materials and methods

Soil samples and DNA extraction

Soil samples were collected from four sites around Scotland (UK). Two sandy loam soil samples, GcA and GcB, with pHs of 4.5 and 7.5, respectively, were collected from a pH-controlled agricultural soil (Scottish Agricultural College, Craibstone, Scotland) that had been maintained for >60 years [25]. A humus-iron podzol soil sample, GcL, was collected from an acidic marsh and a peaty gleyed podzol soil sample, GcC, was collected from an acidic moorland (Table S1). DNA was extracted from 0.5 g of soil as described previously [26]. Three DNA extractions were performed per soil sample, pooled, and cleaned using the DNeasy Power CleanUp Kit according to the manufacturer’s instructions (Qiagen). The resulting DNA quality was assessed by spectrometry (Nanodrop ND-1000) and with 1% agarose gel electrophoresis.

GC fractionation

DNA was then fractionated based on its GC-content using a modified version of a published protocol [18, 27]. Specifically, bisbenzimide (Hoechst 33258) (1 mg/ml) was added to a CsCl solution (refractive index = 1.3970, containing 5 mM Tris, pH 8.0) in a 1:500 (v/v) ratio. One μg of high-quality DNA was added to an 8 ml ultracentrifuge tube (Quick-Seal® Bell-Top Polypropylene Tube, Beckman) before filling the tube with CsCl solution and mixing thoroughly. The ultracentrifuge tubes were then balanced by pair, sealed, and subjected to centrifugation (45 000 rpm (equivalent to 98 318 average g-force) at 4°C for 56 h) in an Optima MAX-XP ultracentrifuge using an MLN-80 rotor (Beckman Coulter). The resulting gradient in the ultracentrifuge tube was fractionated into 16 fractions of 500 μl each using a fraction recovery system (Beckman Coulter). Each fraction’s refractive index was measured using an AR200 Refractometer, Reichert™. Each fractionated sample was mixed with an equal volume of isopropanol followed by a brief centrifugation (16 000 g for 1 min) before discarding the top aqueous layer. This washing procedure was repeated two times to remove excessive bisbenzimide [18]. Each sample was then mixed with twice the volume of PEG solution (30% PEG 6000 and 1.6 M NaCl) and 1 μl glycogen at 4°C overnight, followed by pelleting the DNA with a 16 000 g centrifugation at 4°C for 45 min. The aqueous phase was then discarded, and the DNA pellet was washed with 1 ml of 70% ice-cold ethanol. After a 16 000 g centrifugation at 4°C for 15 min, the ethanol was discarded, and the DNA pellet was dried at 55°C for 1 min before being eluted in 30 μl of molecular biology grade water. All DNA samples were stored at −80°C before further use.

Modelling the average GC-content from the refractive index using pure cultures of bacteria

Three bacterial pure cultures with known GC-content were included, namely Streptococcus uberis NCIMB 702055 (GC-content = 36.6%), Escherichia coli K12 (GC-content = 50.6%) and Micrococcus luteus NCIMB 9278 (GC-content = 73%) (Table S2). DNA extraction of 2-ml fully grown cultures was performed as described previously [28] for the three bacterial pure cultures.

The high-quality DNA from the bacterial pure cultures was fractionated using the same protocol described above. The abundance of the 16S rRNA gene was assessed in each fraction by qPCR using bacterial 16S rRNA primers 341F / 907R (see above). A linear model was tested between the DNA GC-content and the refractive index of bisbenzimide CsCl fraction (Fig. 2A and B, Table S2), as described previously [16]. A strong relationship (R2 = 0.99) existed between the GC-content and the refractive index following equation 1:

Figure 2.

Figure 2

Open in a new tab

Molecular separation of DNA based on GC-content using a bisbenzimide-CsCl DNA fractionation approach. (A) The distribution of the DNA from eight bacterial and three archaeal species across the fractions was determined by qPCR and their refractive index of the fractions was determined. The distribution of DNA from three bacterial species is shown for reference. (B) The refractive index of fractions possessing the DNA of eight bacteria and three archaea was plotted against the known GC-content of their genomes. The relationship between the observed refractive index and their GC-content was determined as per the presented equation. The three colored dots (grey, red, and black) correspond to the bacteria represented in panel A for reference. (C) Amount of DNA across GC-content fractions of the four soils (GcA, GcB, GcC, and GcL). (D) Relative abundance of the bacterial 16S rRNA and archaeal amoA genes across GC-content fractions of four soils (GcA, GcB, GcC, and GcL), estimated based on the 16S rRNA and archaeal amoA gene qPCR on each fraction, respectively. For each soil, 100% represents the total number of each marker gene across all GC-content fractions.

graphic file with name DmEquation1.gif (1)

Refining GC-content prediction with a range of bacterial and archaeal pure cultures

The accuracy of the approach was tested using eight bacterial and three archaeal pure cultures, which span a wide range of GC-content (Table S2). In addition to the three previously described bacteria, five other bacteria were included: Nitrosomonas europea (GC-content = 50.7%), Bacillus subtilis 168 Δ6 (GC-content = 43.3%), Nitrosospira multiformis (GC-content = 53.9%), Pseudomonas putida F1 (GC-content = 62.0%) and P. putida U NCIMB10015 (GC-content = 60.0%). The three selected archaea included Nitrososphaera viennensis (GC-content = 52.7%), Nitrosopumilus maritimus (GC-content = 34.2%) and Nitrosotalea devaniterrae Nd2 (GC-content = 37.4%). Culture media and DNA extraction approaches are described in Table S2. The fractionation procedure used the same protocol as described above.

Another qPCR approach was used to recover both archaea and bacteria, with the universal 16S rRNA primers 515mF (5′- GTG YCA GCM GCC GCG GTA A -3′) [29] and 806mR (5′- GGA CTA CNV GGG TWT CTA AT-3′) [30] due to the enzyme previously used being discontinued by the manufacturer. Here, the 16S rRNA gene qPCR reaction consisted of an initial step at 95°C for 3 min, followed by 40 cycles of 95°C for 15 s, 53°C for 35 s, 72°C for 45 s, and 80°C for 8 s followed by reading, ending with a melting curve analysis. Each 20-μl qPCR reaction consisted of 10 μl iQ SYBR Green™ qPCR master mix (Bio-Rad), 0.2 mg ml−1 bovine serum albumin (BSA), 2 μl DNA template, and 0.5 μM of primers. All qPCR assays used the same full-length E. coli 16S rRNA gene standard as above.

Another linear model was tested between the DNA GC-content and the refractive index of bisbenzimide CsCl fraction and this model was compared to the three-strain one. Both models were highly correlated (see Fig. S1), so a final linear model (equation 2) was used to estimate the average DNA GC-content in each fraction from the soil DNA, based on its refraction index.

graphic file with name DmEquation2.gif (2)

Microbial quantification by qPCR

For the soil DNA, the bacterial 16S rRNA gene and the archaeal amoA gene were quantified in the unfractionated and in each fractionated sample (Table S3). Each 20-μl qPCR reaction consisted of 10 μl QuantiFast™ qPCR master mix (Qiagen), 0.2 mg ml−1 bovine serum albumin (BSA), 2 μl DNA template, and the 0.4 to 0.6 μM of primers. The bacterial 16S rRNA gene primers were 341F / 907R [31] and the archaeal amoA gene primers were 23F/616R [32]. The 16S rRNA gene qPCR reaction consisted of an initial step at 95°C for 5 min, followed by 38 cycles of 94°C for 15 s, 67°C for 45 s, and 80°C for 8 s followed by reading, ending with a melting curve analysis. The archaeal amoA gene qPCR reaction was as described previously [33]. All qPCR assays used a 10-fold serial dilution from 10 to 108 copies of known DNA template as standard. The standard for bacterial 16S rRNA gene qPCR was a full-length E. coli 16S rRNA gene PCR product. A cloned archaeal amoA gene was used as the standard as reported previously [33]. All qPCR assays had R2 values 0.98–1, and their efficiencies were between 95% and 101%. While the 16S rRNA primers targeted the 16S rRNA gene of bacteria, the archaeal amoA primers were specific to AOA (not encompassing the ammonia-oxidising bacterial amoA gene).

Analysis of 16S rRNA gene amplicon library

For universal 16S rRNA gene amplicon sequencing, triplicate PCR reactions targeting the V4 region were performed per sample. Each 25-μl PCR reaction contained 12.5 μl KAPA HiFi HotStart ReadyMix (Roche), 2 μl formamide, 0.5 μM each of the primers 515mF and 806mR with MiSeq sequencing adaptors, and 1 μl DNA sample as template. The PCR amplification was 95°C for 3 min, followed by 35 cycles of 98°C for 20 s, 50°C for 10 s and 72°C for 15 s, and then a final extension at 72°C for 5 min. The products of triplicate PCR were checked on 1% agarose gel and pooled. The 16S rRNA gene amplicon libraries were performed with the Herculase II Fusion DNA Polymerase Nextera XT Index V2 Kit and were sequenced on MiSeq 300PE by Macrogen company to a depth of 179 356 reads per sample on average.

The 16S rRNA gene amplicon sequences were trimmed with TrimGalore v0.6.6 [34] for trimming of sequencing primers and adapters and FastQC v0.11.9 [35] for quality control. Trimmed reads were merged and assigned to amplicon sequence variants (ASVs) using DADA2 v1.30.0 [36], with quality filtering and chimera removal using the default settings (Table S4). ASV libraries were subsampled to 15 735 reads, which was the read number in the smallest library (Table S4) and singletons were removed. Taxonomy (Table S5) was assigned to ASVs in DADA2 using the Genome Taxonomy Database (GTDB) 16S rRNA database (release R207) -(https://data.gtdb.ecogenomic.org/releases/). Community diversity (Table S6) and taxonomic assignment were calculated using the R package, phyloseq v1.46.0 [37] and these outputs were used to estimate the relative abundance of each family. The relationship between communities was calculated using UniFrac, which compares community composition taking into account phylogenetic relatedness of the ASVs. A Principal Coordinate Analysis (PCoA) plot of the community dissimilarity was generated using the phyloseq package.

Metagenomic sequencing and genome assembly

DNA libraries were prepared for several selected high and low GC-content fractions (GcA5, GcA13, GcB4, GcB14, GcC5, GcC11, GcL4, and GcL13) using the Nextera XT DNA Library Preparation Kit and sequenced using Illumina NovaSeq 6000 S4 platform (276 million reads per fraction on average, Table S7) by Macrogen company.

Reads were filtered using the READ_QC module [38], and high-quality reads for each metagenome were assembled using MEGAHIT v1.1.3 [39]. Binning of resulting contigs was performed with MaxBin2 [40] and metaBAT2 [41], and the results were consolidated using the Bin_refinement module [38]. Completeness and contamination of bins were estimated with CheckM v1.1.2 [42], and only bins with a completeness of >45% and contamination of <10% were retained for further analysis (Table S8). Bins were initially characterized using the classify_wf function in GTDB-Tk v2.1.1 [43] using the R207v2 GTDB release. The taxonomy described throughout is based on the GTDB taxonomy and may differ from other sources, such as NCBI taxonomy. Genome coverage and MAG relative abundance were calculated using read recruitment with CoverM v0.6.1 [44] in the fractions from which they were assembled. The relative abundances of each genome’s taxonomic family were also estimated in the unfractionated samples and in each fraction using the 16S rRNA community analysis described above.

Genome characterisation and metabolic prediction

The species tree of the new genomes was created based on the taxonomy provided by GTDB [45]. Novelty of each branch of the tree (defined as the percentage of the total number of genome sequences available for a given taxa that were sequenced in this work) was calculated using the R207v2 GTDB release as a proxy for the diversity of publicly available genomes. Genomes were annotated using Prokka v1.14 [46] and the KEGG database [47] using GhostKOALA [48]. Carbohydrate active enzymes were annotated using profile HMM database dbCAN-HMMdb-V11 [49], filtered with hmmscan-parser.sh and by removing matches with mean posterior probability <0.7. Biosynthetic gene clusters (BGCs) were annotated with antiSMASH v7.1.0 using default settings.

Results

Differential microbial distribution across a GC-content gradient

DNA of eight bacteria and three archaea of known genomic GC-content was fractionated using the GC-fractionation procedure (Fig. 2A) and a linear model (Fig. 2B) demonstrated the strong relationship between the GC-content and the refractive index of the fractions. Based on this relationship, the GC-content was estimated for each fraction of the environmental DNA extracted from four UK soil sites (Craibstone pH 4.5 (GcA) and pH 7.5 (GcB), Corsemaul (GcC), and Lower Dell (GcL)) (Table S1) following the GC-fractionation procedure. The DNA was unevenly distributed across the GC gradient for the four soil samples, with fractions with the highest DNA amounts corresponding to GC-contents of 58%–65% for the four soils (Fig. 2C, Table S3). Most of the DNA was present within 5% GC-content from the peak, and less than 12% of total DNA was present in any fraction with a GC-content <55%.

The abundance of the bacterial 16S rRNA and ammonia monooxygenase subunit A (amoA) genes were determined in each fractionated DNA sample by target-specific qPCR assays (Table S3). While ubiquitous in many environments, AOA were chosen as rare microbes having low GC-content [7]. Within each unfractionated soil DNA, bacteria were between two and four orders of magnitude more abundant than AOA (Table S3). Bacteria and AOA were differently distributed across the GC-gradient in the four soils, with most bacteria (66%–88%) being present in the GC-content range of 50%–70%, while most of the AOA (93%–100%) were present in the low GC-content range (GC < 50%) (Fig. 2D, Table S3). Therefore, AOA were relatively enriched in the low GC-content fractions compared to the high GC-content fractions where bacteria dominated. Bacterial DNA was concentrated within a single range for the soils GcC and GcL, forming a single peak, but appears to form two peaks on the GC-content gradient for the agricultural soils GcA and GcB (Fig. 2D).

Universal 16S rRNA gene sequencing was performed (MiSeq V3 sequencing, Illumina) on each fractioned and unfractionated DNA sample of the four soils (Table S4). Even at the phylum level, there was an observable difference in the microbial composition of each soil along the GC-content gradient. Fractions from both extremes of the gradient possess different communities from the original soils (Fig. 3A, Table S5), and archaea represent at higher relative abundance in the low-GC fractions of three soils (Fig. S2). These results also indicate that the two peaks observed for GcA and GcB in Fig. 2D are the result of the fractionation procedure separating bacterial phyla that are dominant in the unfractionated samples, notably Pseudomonadota and Actinomycetota. There was a trend for higher diversity in the lower GC-content fractions (Fig. 3A, Table S6), although most of the DNA was present in the 50%–70% GC-content range (Fig. 2C). Therefore, low GC fractions are likely an enriched source of DNA for rare microbes. A PCoA plot of the community dissimilarity also indicates compositional change between unfractionated and fractionated samples (Fig. 3B). GC-fractionated pseudo-communities from each site are dissimilar to each other, but are more similar to their original communities rather than to pseudo-communities of other sites with similar GC-content (Fig. 3B).

Figure 3.

Figure 3

Open in a new tab

Microbial communities across a GC-content gradient following a bisbenzimide-CsCl GC-content-based DNA fractionation approach in four soils (GcA, GcB, GcC, and GcL). (A) Taxonomic composition, richness and diversity of microbial community in each unfractionated soil DNA and associated fractionated-based DNA. Richness is defined as the total number of ASVs detected. GC-content percentage of each fraction is indicated in brackets following the fraction name. (B) Principal coordinate analysis of unweighted Unifrac using the microbial composition (ASV-based) of each fraction. The number next to each dot indicates the fraction number of each sample. (C) Taxonomy of metagenome-assembled genomes recovered from each fraction. Numbers above the bars indicate the number of genomes recovered from each fraction. Numbers below the bars indicate the fraction name, followed by the GC-content percentage of the fraction in brackets.

Novelty of genome-resolved metagenomes of rare microbes

Metagenomic sequencing (NovaSeq (Illumina)) was performed on one high and one low GC-content fraction per soil (Table S7). A total of 204 MAGs (completeness >45% and contamination <10%) could be reconstructed from the eight metagenomes: 122 genomes from high GC-content fractions and 82 genomes from low GC-content fractions (Table S8). The MAGs possessed an average completeness of 67% (range 45%–99%) and an average contamination of 4% (range 0%–10%). On average, each genome was present at 0.34% relative abundance in its fractionated pseudo-community (range 0.07%–2.67%) (Table S8). Notably, 31% of genomes were from rare or very rare microbial families (Table S9). The GC-content of most of these genomes corresponded with the GC-content fractions, with high and low GC-content genomes recovered from high and low GC-content fractions, respectively (Fig. S3, Table S10), indicating that GC-content based separation of DNA was successfully achieved for these soils. There was no significant difference in the quality of genomes (based on criteria of completeness and contamination) recovered from the high and low GC-content fractions (Fig. S4).

The 204 MAGs were taxonomically diverse, representing 27 bacterial and four archaeal phyla (Table S11, Fig. 3C) based on the GTDB classification (release R207v2) [50]. Seventy genomes were from previously uncharacterized genera (Fig. 4A; Table S8), whereas the reconstructed MAGs also increased the representation of 24 known genera by over 50%, and almost all MAGs were from novel species (202 of 204 genomes) (Table S8). Most MAGs (61%) recovered in this work are from families that were enriched in the GC-based fractions compared to original soils (enrichment range: 1.1–534.4 -fold) (Table S12), and this molecular enrichment was statistically significant in rare (P = 3.3 × 10−4) and very rare (P = 2.7 × 10−3) microbial families (Fig. 4B).

Figure 4.

Figure 4

Open in a new tab

Recovered metagenome-assembled genomes (MAGs) from the low- and high-GC-content fractions in the four soils. (A) Taxonomy, novelty and genomic characteristics of 204 recovered metagenome-assembled genomes. The curated genome database GTDB R207v2 is used as a proxy to represent publicly available genomes. Phyla with less than four genomes recovered in this study are classified as “other phyla” in this figure. “No KEGG” indicates the percentage of genes with no close homologs in the KEGG database. Details of each MAG and genome novelty of each branch are available in Tables S8 and S11, respectively. (B) Relative abundance of genome families in the original and fractionated soil DNAs for families being abundant (> 1%), minor (< 1%), rare (< 0.1%), and very rare (< 0.01%) in the original soil DNA. The family taxonomy is assessed based on the genome sequence, while the relative abundances are determined through 16S rRNA gene amplicon sequencing analysis. The number of genome families (n) in each category is indicated. P-values indicate statistical differences between the two groups (t-test 1-tail paired) for each category. A low amount of data within the fractionated samples was omitted from the graphical representation (i.e. one, one, and five values for the minor, rare, and very rare categories, respectively) because they exceeded the maximum value of the y-axis. Detailed family abundance information for each recovered genome is available in Table S9.

The MAGs sequenced in this study possessed a high percentage of genes with no close homologs in the KEGG database. The average percentage of uncharacterized genes per genome was 55% (range 30%–78%) (Table S8). Despite this, genes involved in important biogeochemical cycles and carbohydrate degradation were often detected (Tables S13 and S14). In total, 670 biosynthetic gene clusters (BGCs) were detected in the 204 genomes, with most genomes (76%) possessing at least one BGC. Ribosomally synthesized and post-translationally modified peptides were the most frequently detected category of BGC (32%), followed by polyketide synthases (20%), terpene (19%), non-ribosomal peptides (18%), and other BGCs (10%) (Table S15). There was no significant difference between the number of BGCs in genomes from different abundance ranks. Finally, BGCs were mostly not retrieved in the genomes of certain taxa, such as the Chloroflexota family CSP1–4 and the Patescibacteria phylum (Table S15).

Rare biosphere genomes of the Bacteroidia

As the majority of DNA was present in the higher GC-content fractions, the lower GC-fractions were more likely to contain DNA from rarer microbial families. Notably, 19 genomes sequences were recovered from the low-GC content possessing Bacteroidia class from the orders Chitinophagales (10 genomes), Sphingobacteriales (3 genomes), and Flavobacteriales (1 genome), and the uncharacterized orders o__AKYH767 (3 genomes), o__AKYH767-A (1 genome), and o__NS11–12 g (1 genome). Seven of these genomes were from rare families (< 0.1% relative abundance), and six were from very rare families (< 0.01% relative abundance) that were enriched 3.2 to 56.8 -fold, compared to soils (Table 1).

Table 1.

Recovered Bacteroidia genomes, family enrichment score following GC-fractionation procedure and family abundance rank in the original, unfractionated sample. The genome completeness, genome contamination scores as well as taxonomy (order, family, and genus) are indicated for each MAG. In addition, the relative abundance of each genome’s family in the unfractionated sample and in the fraction of interest are indicated. For the family abundance rank, each family was classified as abundant (> 1%), minor (0.1%–1%), rare (0.01%–0.1%), or very rare (< 0.01%), based on its 16S rRNA gene relative abundance in the unfractionated sample. For each genome’s family, the amount of enrichment resulting from the GC fractionation is represented by the ratio of the 16S rRNA gene relative abundance in the fraction of interest compared to the unfractionated sample.

MAG name Completeness (%) Contamination (%) Order Family Genus Relative abundance in unfractionated DNA (%) Abundance rank Relative abundance in fraction of interest (%) Fold family enrichment
GcB14–4 87 3 o__AKYH767-A f__OLB10 g__ 0.09 Rare 0.28 3
GcL13–5 66 0 o__NS11–12 g f__UKL13–3 g__CAIUEQ01 0.00 Very rare 0.08 *
GcB14–6 52 3 o__Flavobacteriales f__UBA2798 g__ 0.00 Very rare 0.03 *
GcA13–2 81 2 o__Sphingobacteriales f__Sphingobacteriaceae g__Mucilaginibacter 0.55 Minor 3.17 6
GcA13–16 62 5 o__Sphingobacteriales f__Sphingobacteriaceae g__Mucilaginibacter 0.55 Minor 3.17 6
GcC11–12 83 3 o__Sphingobacteriales f__Sphingobacteriaceae g__Mucilaginibacter 0.02 Rare 1.31 57
GcC11–29 57 2 o__AKYH767 f__B-17BO g__PALSA-968 0.10 Rare 4.51 48
GcC11–26 99 0 o__AKYH767 f__Palsa-965 g__Palsa-965 0.00 Very rare 0.26 *
GcL13–16 48 1 o__AKYH767 f__Palsa-948 g__Palsa-948 0.01 Very rare 0.26 51
GcB14–5 54 0 o__Chitinophagales f__BACL12 g__UBA7236 0.00 Very rare 0.07 *
GcL13–2 48 2 o__Chitinophagales f__UBA10324 g__JAFDYS01 0.00 Very rare 0.01 *
GcB14–9 89 4 o__Chitinophagales f__Chitinophagaceae g__VBAT01 0.91 Minor 5.21 6
GcC11–6 51 0 o__Chitinophagales f__Chitinophagaceae g__Flavipsychrobacter 0.39 Minor 0.51 1
GcL13–3 46 3 o__Chitinophagales f__Chitinophagaceae g__UTBCD1 0.09 Rare 3.52 40
GcL13–12 87 0 o__Chitinophagales f__Chitinophagaceae g__UTBCD1 0.09 Rare 3.52 40
GcL13–14 55 7 o__Chitinophagales f__Chitinophagaceae g__Puia 0.09 Rare 3.52 40
GcA13–14 81 4 o__Chitinophagales f__Chitinophagaceae g__Puia 1.01 Abundant 17.10 17
GcA13–9 59 8 o__Chitinophagales f__Chitinophagaceae g__Puia 1.01 Abundant 17.10 17
GcL13–9 70 6 o__Chitinophagales f__Chitinophagaceae g__Puia 0.09 Rare 3.52 40
Open in a new tab

*Family was detected in fraction, but was too low abundance to be detected in original soil.

As noted for the Bacteroidia families Bacteroidaceae and Flavobacteriaceae, these microbes appear to not form endospores [51, 52], as they lack genes for endospore formation (Table S16). All Bacteroidia recovered in this study are predicted to be non-motile as they lacked genes for chemotaxis and flagellar assembly (Table S16). Most Bacteroidia genomes possess marker genes for aerobic respiration (coxA, coxB, ctaA, and ctaB), and several genomes also have genes for microaerobic respiration terminal oxidases, cytochrome bd ubiquinol oxidase (cydA and cydB) [53–55], and cytochrome cbb3 oxidase (ccoN, ccoNO, ccoO, and ccoP) [56] (Table S16). Additionally, they lack evident marker genes of a strictly anaerobic lifestyle, such as fumarate hydratase class 1 (fumA and fumB) [57] and pyruvate formate lyase (pflD) (Table S16). Together this indicates that these organisms are adapted to aerobic or microaerobic lifestyles.

Lactate is a key metabolite in microbial cross-feeding in the human gut [58–61] and its utilization significantly influences colonic ecosystem pH, microbial metabolism, and community stability [60]. In contrast, the role of lactate cycling in soils is less well understood, and its contribution to soil microbial ecosystems remains an open question. Most of the Bacteroidia genomes possess a three-component- type L-lactate dehydrogenase [62], critical for converting L-lactate to pyruvate. Only one genome, GcL13–9, possesses a lactate permease, which has been predicted to be necessary for utilising exogenous lactate [63]. Manual curation of the genomic region surrounding the three-component L-lactate dehydrogenase genes in GcA13–16 revealed the presence of an L-fucose permease, and an L-fucose degradation pathway that produces L-lactate as an end product (Table S17). The L-lactate dehydrogenase was present in many of the reconstructed Bacteroidia genomes (Table S17), indicating that this enzyme may, in fact, function in utilising intracellularly produced lactate as part of L-fucose utilisation.

Discussion

The rare biosphere makes up a significant portion of the diversity of our planet [64]. Many studies focus on the more abundant members of microbial communities, but low-abundance microbes also drive some critical ecosystem functions, especially when performing metabolic activities that are taxonomically restricted, such as in the case of microbial ammonia oxidation [65]. While it is difficult to reconstruct genomes from rare environmental microbes with the classical metagenomic approach, we demonstrated the potential of GC-fractionation to overcome this gap. In this proof-of-concept work, 63 genomes were recovered from rare or very rare microbial families. It is unclear whether these low-abundance organisms are naturally ecologically constrained to a low abundance by virtue of their biochemical processes or whether they act as a “functional seed bank” [10], being only transiently in low abundance and increasing in abundance under certain environmental stimuli, such as acidic/osmotic stress or the intermittent introduction of a specific nutrient into the ecosystem. Recent co-occurrence network analysis suggests that low-abundance taxa have stronger ecological relevance to the community than more abundant taxa in terrestrial ecosystems [66], but a mechanistic explanation for this is lacking. Obtaining genome sequences for these rare taxa would enable a greater knowledge of their metabolic potential and role in soil microbial ecosystems. Those low-abundance taxa may represent habitat specialist groups, in contrast to the more abundant microbes being habitat generalists [66–68], even if a wide range of natural abundance has been observed for generalists [69, 70].

More than half of the reconstructed genomes (124 out of 204) belonged to families whose 16S rRNA gene relative abundance was increased from the unfractionated to the fractionated samples. While family enrichments up to 534-fold could be calculated, it is also notable that 20 genomes were assembled from families that were not detected in the corresponding unfractionated soil, preventing calculation of the enrichment of these families. However, based on the number of amplicons sequenced in the unfractionated samples (subsampled to 15 735) and the abundance of these families in the fractionated samples, it appears that many of these genomes belong to families that have been enriched to an even greater degree. Notably, the Dormibacteria family f__UBA8260 was potentially enriched more than 839-fold in the Corsemaul (GcC5) high-GC fraction, compared to the unfractionationed sample (GcC). Taking these genomes into account in the overall family enrichment score, ~60% of the reconstructed genomes belong to families that have been selectively enriched using the described GC-fractionation approach.

The recovered genomes presented extensive taxonomic diversity, especially for the low GC-content MAGs, representing a diverse pool of rare microbes. Among the recovered genomes, several genomes represented novel orders (e.g. GcC11–19, novel order within the Gammaproteobacteria class, and GcA5–32, novel order within the UBA5177 class), novel families (14 novel families), and numerous novel genera (67 genomes belonging to novel genera) (Table S8). Reconstruction of previously undiscovered genomic diversity is critical to advancing evolutionary questions, such as those related to microbial diversification rates [71, 72] or the origins of eukaryotes [73, 74]. In addition, the reconstructed genomes included 19 genomes belonging to the Bacteroidia class, allowing some metabolic inference of this group. Such genome reconstruction could also further clarify the ecological strategy of microbes, as low GC genomes were reported to follow oligotrophic strategies, while high GC genomes are copiotrophs [75]. This demonstrates that this selective sequencing approach enables reconstruction of multiple genomes from specific taxonomic groups. Therefore, such increased genomic representation allows metabolic inference of poorly studied microbial groups and deeper understanding of the mechanisms of microbial evolutionary genomic dynamics [7, 76–78].

Supplementary Material

20250731_GC_fract_SI_text_ycaf152
20250731_gc_fract_si_text_ycaf152.pdf (647.9KB, pdf)
20250610_GC_fract_SI_tables_ycaf152
20250610_gc_fract_si_tables_ycaf152.xlsx (866KB, xlsx)

Acknowledgements

The authors would also like to acknowledge the support of the Maxwell computer cluster funded by the University of Aberdeen. In addition, we acknowledge Scotland’s Rural College (SRUC) as the custodian of the Aberdeen Cropping Experimental (ACE) Platform which hosts both the pHoenix and Tulloch Organic Rotation long-term experiments (LTE’s) and was the source of some of the soil samples used as part of this study. The maintenance of these LTEs is funded, in part, from the Scottish Government’s RESAS programme (Healthy soil for a Green Recovery). For the purpose of open access, the authors have applied a Creative Commons Attribution (CC BY) licence to any Author Accepted Manuscript version arising from this submission.

Contributor Information

Paul O Sheridan, School of Biological Sciences, University of Aberdeen, Aberdeen AB24 3UU, United Kingdom; School of Biological and Chemical Sciences, University of Galway, Galway H91 TK33, Ireland.

Yiyu Meng, School of Biological Sciences, University of Aberdeen, Aberdeen AB24 3UU, United Kingdom.

Dylan Bodington, School of Biological Sciences, University of Aberdeen, Aberdeen AB24 3UU, United Kingdom.

David Coutts, School of Biological Sciences, University of Aberdeen, Aberdeen AB24 3UU, United Kingdom.

Tom A Williams, School of Biological Sciences, University of Bristol, Bristol BS81TQ, United Kingdom.

Cécile Gubry-Rangin, School of Biological Sciences, University of Aberdeen, Aberdeen AB24 3UU, United Kingdom.

Author contributions

C.G.-R. conceptualized the study. Investigation was conducted by P.O.S. and Y.M. Formal analysis was performed by P.O.S. and D.B. P.O.S. and C.G.-R. interpreted the results and wrote the manuscript. All authors have accepted the final version of the manuscript

Conflicts of interest

The authors have nothing to disclose.

Funding

UKRI financially supported P.O.S. and Y.M. through the NERC grant (NE/R001529/1). In addition, C.G.-R. was supported by a Royal Society University Research Fellowship (URF150571) and D.B. received a PhD studentship from the Royal Society (RG160625).

Code and data availability

Scripts for general manipulation of the 16S rRNA gene diversity have been deposited at https://github.com/bodington/gc_16S/, and scripts specific to the genome novelty have been deposited at https://github.com/SheridanPO-Lab/Genome-novelty. The 16S rRNA gene sequencing data and the 204 new genome sequences presented in this work have been deposited under the NCBI BioProject PRJNA1160233.

References

  • 1. Hug  LA, Baker  BJ, Anantharaman  K. et al.  A new view of the tree of life. Nat Microbiol  2016;1:1–6. 10.1038/nmicrobiol.2016.48 [DOI] [PubMed] [Google Scholar]
  • 2. Adam  PS, Borrel  G, Brochier-Armanet  C. et al.  The growing tree of Archaea: new perspectives on their diversity, evolution and ecology. The ISME journal  2017;11:2407–25. 10.1038/ismej.2017.122 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Jousset  A, Bienhold  C, Chatzinotas  A. et al.  Where less may be more: how the rare biosphere pulls ecosystems strings. The ISME journal  2017;11:853–62. 10.1038/ismej.2016.174 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Bars-Cortina  D, Ramon  E, Rius-Sansalvador  B. et al.  Comparison between 16S rRNA and shotgun sequencing in colorectal cancer, advanced colorectal lesions, and healthy human gut microbiota. BMC Genomics  2024;25:730. 10.1186/s12864-024-10621-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Stothart  MR, McLoughlin  PD, Poissant  J. Shallow shotgun sequencing of the microbiome recapitulates 16S amplicon results and provides functional insights. Mol Ecol Resour  2023;23:549–64. 10.1111/1755-0998.13713 [DOI] [PubMed] [Google Scholar]
  • 6. Nicholls  SM, Quick  JC, Tang  S. et al.  Ultra-deep, long-read nanopore sequencing of mock microbial community standards. Gigascience  2019;8:giz043. 10.1093/gigascience/giz043 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Sheridan  PO, Raguideau  S, Quince  C. et al.  Gene duplication drives genome expansion in a major lineage of Thaumarchaeota. Nat Commun  2020;11:5494–12. 10.1038/s41467-020-19132-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Zheng  Y, Wang  B, Gao  P. et al.  Novel order-level lineage of ammonia-oxidizing archaea widespread in marine and terrestrial environments. The ISME Journal  2024;18:wrad002. 10.1093/ismejo/wrad002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Alves  RJE, Minh  BQ, Urich  T. et al.  Unifying the global phylogeny and environmental distribution of ammonia-oxidising archaea based on amoA genes. Nat Commun  2018;9:1–17. 10.1038/s41467-018-03861-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Lennon  JT, Jones  SE. Microbial seed banks: the ecological and evolutionary implications of dormancy. Nat Rev Microbiol  2011;9:119–30. 10.1038/nrmicro2504 [DOI] [PubMed] [Google Scholar]
  • 11. Cross  KL, Campbell  JH, Balachandran  M. et al.  Targeted isolation and cultivation of uncultivated bacteria by reverse genomics. Nat Biotechnol  2019;37:1314–21. 10.1038/s41587-019-0260-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Bowers  RM, Kyrpides  NC, Stepanauskas  R. et al.  Minimum information about a single amplified genome (MISAG) and a metagenome-assembled genome (MIMAG) of bacteria and archaea. Nat Biotechnol  2017;35:725–31. 10.1038/nbt.3893 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Cao  L, Kong  Y, Fan  Y. et al.  mEnrich-seq: methylation-guided enrichment sequencing of bacterial taxa of interest from microbiome. Nat Methods  2024;21:236–46. 10.1038/s41592-023-02125-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Gu  W, Crawford  ED, O’Donovan  BD. et al.  Depletion of abundant sequences by hybridization (DASH): using Cas9 to remove unwanted high-abundance species in sequencing libraries and molecular counting applications. Genome Biol  2016;17:1–13. 10.1186/s13059-016-0904-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Sun  Y, Cheng  Z, Li  X. et al.  Genome enrichment of rare and unknown species from complicated microbiomes by nanopore selective sequencing. Genome Res  2023;33:612–21. 10.1101/gr.277266.122 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Holben  WE, Harris  D. DNA-based monitoring of total bacterial community structure in environmental samples. Mol Ecol  1995;4:627–32. 10.1111/j.1365-294X.1995.tb00263.x [DOI] [PubMed] [Google Scholar]
  • 17. Chuckran  PF, Hungate  BA, Schwartz  E. et al.  Variation in genomic traits of microbial communities among ecosystems. FEMS microbes  2021;2:xtab020. 10.1093/femsmc/xtab020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Tran-Nguyen  LT, Schneider  B. Cesium chloride-bisbenzimide gradients for separation of phytoplasma and plant DNA. In: Dickinson M, Hodgetts, J (eds), Phytoplasma: Methods and Protocols, 2013, 381–93. Humana Press, Totowa, New Jersey, US. 10.1007/978-1-62703-089-2_32 [DOI] [PubMed]
  • 19. Holben  WE, Feris  KP, Kettunen  A. et al.  GC fractionation enhances microbial community diversity assessment and detection of minority populations of bacteria by denaturing gradient gel electrophoresis. Appl Environ Microbiol  2004;70:2263–70. 10.1128/AEM.70.4.2263-2270.2004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Holben  WE, Noto  K, Sumino  T. et al.  Molecular analysis of bacterial communities in a three-compartment granular activated sludge system indicates community-level control by incompatible nitrification processes. Appl Environ Microbiol  1998;64:2528–32. 10.1128/AEM.64.7.2528-2532.1998 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Santo Domingo  JW, Kaufman  MG, Klug  MJ. et al.  Influence of diet on the structure and function of the bacterial hindgut community of crickets. Mol Ecol  1998;7:761–7. 10.1046/j.1365-294x.1998.00390.x [DOI] [Google Scholar]
  • 22. Fuhrman  JA. Microbial community structure and its functional implications. Nature  2009;459:193–9. 10.1038/nature08058 [DOI] [PubMed] [Google Scholar]
  • 23. Galand  PE, Casamayor  EO, Kirchman  DL. et al.  Ecology of the rare microbial biosphere of the Arctic Ocean. Proc Natl Acad Sci  2009;106:22427–32. 10.1073/pnas.0908284106 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Bardgett  RD, van der Putten  W, Der Putten  V. et al.  Belowground biodiversity and ecosystem functioning. Nature  2014;515:505–11. 10.1038/nature13855 [DOI] [PubMed] [Google Scholar]
  • 25. Kemp  JS, Paterson  E, Gammack  SM. et al.  Leaching of genetically modified Pseudomonas fluorescens through organic soils: influence of temperature, soil pH, and roots. Biol Fertility Soils  1992;13:218–24. 10.1007/BF00340579 [DOI] [Google Scholar]
  • 26. Nicol  GW, Tscherko  D, Embley  TM. et al.  Primary succession of soil Crenarchaeota across a receding glacier foreland. Environ Microbiol  2005;7:337–47. 10.1111/j.1462-2920.2005.00698.x [DOI] [PubMed] [Google Scholar]
  • 27. Holben  WE. GC fractionation allows comparative total microbial community analysis, enhances diversity assessment, and facilitates detection of minority populations of bacteria. In: de Bruijn, FJ (eds), Handbook of Molecular Microbial Ecology I: Metagenomics and Complementary Approaches. 2011, 183–96. Wiley, Hoboken, New Jersey, US. 10.1002/9781118010518.ch23 [DOI] [Google Scholar]
  • 28. Tourna  M, Freitag  TE, Prosser  JI. Stable isotope probing analysis of interactions between ammonia oxidizers. Appl Environ Microbiol  2010;76:2468–77. 10.1128/AEM.01964-09 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Parada  AE, Needham  DM, Fuhrman  JA. Every base matters: assessing small subunit rRNA primers for marine microbiomes with mock communities, time series and global field samples. Environ Microbiol  2016;18:1403–14. 10.1111/1462-2920.13023 [DOI] [PubMed] [Google Scholar]
  • 30. Apprill  A, McNally  S, Parsons  R. et al.  Minor revision to V4 region SSU rRNA 806R gene primer greatly increases detection of SAR11 bacterioplankton. Aquat Microb Ecol  2015;75:129–37. 10.3354/ame01753 [DOI] [Google Scholar]
  • 31. Muyzer  G, De Waal  EC, Uitterlinden  A. Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA. Appl Environ Microbiol  1993;59:695–700. 10.1128/aem.59.3.695-700.1993 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Tourna  M, Freitag  TE, Nicol  GW. et al.  Growth, activity and temperature responses of ammonia-oxidizing archaea and bacteria in soil microcosms. Environ Microbiol  2008;10:1357–64. 10.1111/j.1462-2920.2007.01563.x [DOI] [PubMed] [Google Scholar]
  • 33. Thion  C, Prosser  JI. Differential response of nonadapted ammonia-oxidising archaea and bacteria to drying–rewetting stress. FEMS Microbiol Ecol  2014;90:380–9. 10.1111/1574-6941.12395 [DOI] [PubMed] [Google Scholar]
  • 34. Krueger  F. Trim Galore: a wrapper tool around Cutadapt and FastQC to consistently apply quality and adapter trimming to FastQ files, with some extra functionality for MspI-digested RRBS-type (Reduced Representation Bisufite-Seq) libraries. http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/. (Date of access: 28/04/2016) 2012.
  • 35. Andrews  S. FastQC: a quality control tool for high throughput sequence data. 2010. https://www bioinformatics babraham ac uk/projects/fastqc/
  • 36. Callahan  BJ, McMurdie  PJ, Rosen  MJ. et al.  DADA2: high-resolution sample inference from Illumina amplicon data. Nat Methods  2016;13:581–3. 10.1038/nmeth.3869 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. McMurdie  PJ, Holmes  S. Phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One  2013;8:e61217. 10.1371/journal.pone.0061217 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Uritskiy  GV, DiRuggiero  J, Taylor  J. MetaWRAP—a flexible pipeline for genome-resolved metagenomic data analysis. Microbiome  2018;6:1–13. 10.1186/s40168-018-0541-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Li  D, Liu  C, Luo  R. et al.  MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics  2015;31:1674–6. 10.1093/bioinformatics/btv033 [DOI] [PubMed] [Google Scholar]
  • 40. Wu  Y, Simmons  BA, Singer  SW. MaxBin 2.0: an automated binning algorithm to recover genomes from multiple metagenomic datasets. Bioinformatics  2016;32:605–7. 10.1093/bioinformatics/btv638 [DOI] [PubMed] [Google Scholar]
  • 41. Kang  DD, Li  F, Kirton  E. et al.  MetaBAT 2: an adaptive binning algorithm for robust and efficient genome reconstruction from metagenome assemblies. PeerJ  2019;7:e7359. 10.7717/peerj.7359 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Parks  DH, Imelfort  M, Skennerton  CT. et al.  CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Res  2015;25:1043–55. 10.1101/gr.186072.114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43. Chaumeil  P, Mussig  AJ, Hugenholtz  P. et al.  GTDB-Tk: a toolkit to classify genomes with the genome taxonomy database. Bioinformatics  2018;36:1925–7. 10.1093/bioinformatics/btz848 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. Aroney  ST, Newell  RJ, Nissen  JN. et al.  CoverM: read alignment statistics for metagenomics. Bioinformatics  2025;41:btaf147. 10.1093/bioinformatics/btaf147 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45. Parks  DH, Chuvochina  M, Rinke  C. et al.  GTDB: an ongoing census of bacterial and archaeal diversity through a phylogenetically consistent, rank normalized and complete genome-based taxonomy. Nucleic Acids Res  2022;50:D785–94. 10.1093/nar/gkab776 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46. Seemann  T. Prokka: rapid prokaryotic genome annotation. Bioinformatics  2014;30:2068–9. 10.1093/bioinformatics/btu153 [DOI] [PubMed] [Google Scholar]
  • 47. The KEGG database . Novartis Foundation Symposium. Wiley Online Library, 2002. Wiley, Hoboken, New Jersey, US. [Google Scholar]
  • 48. Kanehisa  M, Sato  Y, Morishima  K. BlastKOALA and GhostKOALA: KEGG tools for functional characterization of genome and metagenome sequences. J Mol Biol  2016;428:726–31. 10.1016/j.jmb.2015.11.006 [DOI] [PubMed] [Google Scholar]
  • 49. Yin  Y, Mao  X, Yang  J. et al.  dbCAN: a web resource for automated carbohydrate-active enzyme annotation. Nucleic Acids Res  2012;40:W445–51. 10.1093/nar/gks479 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50. Parks  DH, Chuvochina  M, Waite  DW. et al.  A standardized bacterial taxonomy based on genome phylogeny substantially revises the tree of life. Nat Biotechnol  2018;36:996–1004. 10.1038/nbt.4229 [DOI] [PubMed] [Google Scholar]
  • 51. Bernardet  J, Segers  P, Vancanneyt  M. et al.  Cutting a Gordian knot: emended classification and description of the genus Flavobacterium, emended description of the family Flavobacteriaceae, and proposal of Flavobacterium hydatis nom. nov. (basonym, Cytophaga aquatilis Strohl and Tait 1978). Int J Syst Evol Microbiol  1996;46:128–48. [Google Scholar]
  • 52. Krieg  NR, Order  I. Bacteroidales ord. nov. In: Whitman, WB (eds), Bergey's Manual of Systematic Bacteriology: Volume 4: The Bacteroidetes, Spirochaetes, Tenericutes (Mollicutes), Acidobacteria, Fibrobacteres, Fusobacteria, Dictyoglomi, Gemmatimonadetes, Lentisphaerae, Verrucomicrobia, Chlamydiae, and Planctomycetes, Vol. 4, 2011, 25. Springer, New York, US. [Google Scholar]
  • 53. Rice  CW, Hempfling  WP. Oxygen-limited continuous culture and respiratory energy conservation in Escherichia coli. J Bacteriol  1978;134:115–24. 10.1128/jb.134.1.115-124.1978 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54. Tseng  C, Albrecht  J, Gunsalus  RP. Effect of microaerophilic cell growth conditions on expression of the aerobic (cyoABCDE and cydAB) and anaerobic (narGHJI, frdABCD, and dmsABC) respiratory pathway genes in Escherichia coli. J Bacteriol  1996;178:1094–8. 10.1128/jb.178.4.1094-1098.1996 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55. Baughn  AD, Malamy  MH. The strict anaerobe Bacteroides fragilis grows in and benefits from nanomolar concentrations of oxygen. Nature  2004;427:441–4. 10.1038/nature02285 [DOI] [PubMed] [Google Scholar]
  • 56. Pitcher  RS, Brittain  T, Watmugh  NJ. Cytochrome cbb 3 oxidase and bacterial microaerobic metabolism. Biochem Soc Trans  2002;30:653–8. 10.1042/bst0300653 [DOI] [PubMed] [Google Scholar]
  • 57. Davin  AA, Woodcroft  BJ, Soo  RM. et al.  A geological timescale for bacterial evolution and oxygen adaptation. Science  2025;388: eadp1853. 10.1126/science.adp1853 [DOI] [PubMed] [Google Scholar]
  • 58. Louis  P, Duncan  SH, Sheridan  PO. et al.  Microbial lactate utilisation and the stability of the gut microbiome. Gut Microbiome  2022;3:e3. 10.1017/gmb.2022.3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59. Zhao  S, Lau  R, Zhong  Y. et al.  Lactate cross-feeding between Bifidobacterium species and Megasphaera indica contributes to butyrate formation in the human colonic environment. Appl Environ Microbiol  2024;90:1019. 10.1128/aem.01019-23 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60. Wang  SP, Rubio  LA, Duncan  SH. et al.  Pivotal roles for pH, lactate, and lactate-utilizing bacteria in the stability of a human colonic microbial ecosystem. Msystems  2020;5:00645–20. 10.1128/mSystems.00645-20 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61. Hoek MJv, Merks RM.  Emergence of microbial diversity due to cross-feeding interactions in a spatial model of gut microbial metabolism. BMC Syst Biol  2017;11:56. 10.1186/s12918-017-0430-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62. Pinchuk  GE, Rodionov  DA, Yang  C. et al.  Genomic reconstruction of Shewanella oneidensis MR-1 metabolism reveals a previously uncharacterized machinery for lactate utilization. Proc Natl Acad Sci USA  2009;106:2874–9. 10.1073/pnas.0806798106 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63. Sheridan  PO, Louis  P, Tsompanidou  E. et al.  Distribution, organization and expression of genes concerned with anaerobic lactate utilization in human intestinal bacteria. Microbial genomics  2022;8:000739. 10.1099/mgen.0.000739 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64. Nemergut  DR, Costello  EK, Hamady  M. et al.  Global patterns in the biogeography of bacterial taxa. Environ Microbiol  2011;13:135–44. 10.1111/j.1462-2920.2010.02315.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 65. Abell  GC, Robert  SS, Frampton  DM. et al.  High-throughput analysis of ammonia oxidiser community composition via a novel, amoA-based functional gene array. PLoS One  2012;7:e51542. 10.1371/journal.pone.0051542 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 66. Riddley  M, Hepp  S, Hardeep  F. et al.  Differential roles of deterministic and stochastic processes in structuring soil bacterial ecotypes across terrestrial ecosystems. Nat Commun  2025;16:2337. 10.1038/s41467-025-57526-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 67. Székely  AJ, Langenheder  S. The importance of species sorting differs between habitat generalists and specialists in bacterial communities. FEMS Microbiol Ecol  2014;87:102–12. 10.1111/1574-6941.12195 [DOI] [PubMed] [Google Scholar]
  • 68. von Meijenfeldt  FB, Hogeweg  P, Dutilh  BE. A social niche breadth score reveals niche range strategies of generalists and specialists. Nature Ecology & Evolution  2023;7:768–81. 10.1038/s41559-023-02027-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 69. Gubry-Rangin  C, Aigle  A, Herrera-Alsina  L. et al.  Niche breadth specialization impacts ecological and evolutionary adaptation following environmental change. The ISME Journal  2024;18:wrae183. 10.1093/ismejo/wrae183 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 70. Pandit  SN, Kolasa  J, Cottenie  K. Contrasts between habitat generalists and specialists: an empirical extension to the basic metacommunity framework. Ecology  2009;90:2253–62. 10.1890/08-0851.1 [DOI] [PubMed] [Google Scholar]
  • 71. Gubry-Rangin  C, Kratsch  C, Williams  TA. et al.  Coupling of diversification and pH adaptation during the evolution of terrestrial Thaumarchaeota. Proc Natl Acad Sci  2015;112:9370–5. 10.1073/pnas.1419329112 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 72. Feng  W, Wan  X, Zhang  Y. et al.  Diversification, niche adaptation, and evolution of a candidate phylum thriving in the deep critical zone. Proc Natl Acad Sci  2025;122:e2424463122. 10.1073/pnas.2424463122 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 73. Eme  L, Tamarit  D, Caceres  EF. et al.  Inference and reconstruction of the heimdallarchaeial ancestry of eukaryotes. Nature  2023;618:992–9. 10.1038/s41586-023-06186-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 74. Donoghue  PC, Kay  C, Spang  A. et al.  Defining eukaryotes to dissect eukaryogenesis. Curr Biol  2023;33:R919–29. 10.1016/j.cub.2023.07.048 [DOI] [PubMed] [Google Scholar]
  • 75. Okie  JG, Poret-Peterson  AT, Lee  ZM. et al.  Genomic adaptations in information processing underpin trophic strategy in a whole-ecosystem nutrient enrichment experiment. Elife  2020;9:e49816. 10.7554/eLife.49816 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 76. Sheridan  PO, Meng  Y, Williams  TA. et al.  Genomics of soil depth niche partitioning in the Thaumarchaeota family Gagatemarchaeaceae. Nat Commun  2023;14:7305. 10.1038/s41467-023-43196-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 77. Sheridan  PO, Meng  Y, Williams  TA. et al.  Recovery of Lutacidiplasmatales archaeal order genomes suggests convergent evolution in Thermoplasmatota. Nat Commun  2022;13:1–13. 10.1038/s41467-022-31847-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 78. Williams  TA, Davin  AA, Szánthó  LL. et al.  Phylogenetic reconciliation: making the most of genomes to understand microbial ecology and evolution. The ISME Journal  2024;18:wrae129. 10.1093/ismejo/wrae129 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

20250731_GC_fract_SI_text_ycaf152
20250731_gc_fract_si_text_ycaf152.pdf (647.9KB, pdf)
20250610_GC_fract_SI_tables_ycaf152
20250610_gc_fract_si_tables_ycaf152.xlsx (866KB, xlsx)

Data Availability Statement

Scripts for general manipulation of the 16S rRNA gene diversity have been deposited at https://github.com/bodington/gc_16S/, and scripts specific to the genome novelty have been deposited at https://github.com/SheridanPO-Lab/Genome-novelty. The 16S rRNA gene sequencing data and the 204 new genome sequences presented in this work have been deposited under the NCBI BioProject PRJNA1160233.

Articles from ISME Communications are provided here courtesy of Oxford University Press

RESOURCES