A Resource Benchmark of Human Microbiota–Associated Mouse Models Reveals Donor-Dominated Outcomes

Updated on June 05, 2026

Contents Introduction Analysis-ready dataset Methods Results Pre-processing Microbiome Discussion Data Availability Code Availability References Abbreviations

Color key

Introduction

• Background
• HMA rodent models are widely used for causal microbiome-phenotype studies, but the ~95% reported phenotype-transfer rate overstates microbiome causality, and donor taxa frequently fail to colonize or diverge from donor profiles in recipients ^{Walter et al. (2020)}.
• MICROBENCH addresses this with a longitudinal metagenomic benchmarking resource: 10 donors × germ-free C57BL/6 recipients, profiled by long-read shotgun metagenomics for bacterial community composition and mycobiome/virome fractions across fecal and intestinal (Colon, Ileum, Cecum) samples. This report covers the metagenomics component; the companion metabolomic and immune analyses are reported separately.
• Rationale
• Donor-to-recipient transfer fidelity is rarely quantified rigorously across donors, time points, and gut compartments within a single experiment; existing HMA benchmarks are typically limited in donor number or sampling breadth.
• A longitudinal metagenomic dataset spanning multiple donors, time points, and gut compartments is therefore needed as a community-level benchmark for HMA studies and FMT trial design.
• Objective
• Deliver MICROBENCH and a donor-tailored long-read metagenomics pipeline as a calibrated reference for future HMA cohorts and FMT trial designs.




• Aims of the study
• Generate a longitudinal metagenomic benchmarking dataset of germ-free C57BL/6 mice colonized by fecal microbiota transfer from 10 healthy human donors, profiled for bacterial community composition and mycobiome/virome fractions across fecal and intestinal (Colon, Ileum, Cecum) samples.
• Quantify donor-to-recipient transfer fidelity across fecal and intestinal compartments and over the post-FMT time course (Weeks 1–8).
• Establish and release an end-to-end long-read metagenomics pipeline for HMA studies—comprising donor-tailored reference database construction, read-level compositional quality control, and genome-size-aware species-level abundance estimation—so that the dataset and pipeline together yield reference statistics for donor-to-recipient microbiota transfer that can be used to benchmark future HMA studies and support FMT trial design.

Analysis-ready dataset

1. Donors, N=10
1. Eight female (F) and two male (M) healthy donors


2. Samples (post-QC), N=1,249
1. Donor, N=10
2. Donor replicates, N=9 (F1 not replicated)
3. OMM12, N=12
4. Neg. control, N=8
5. Mouse, N=1,210


3. Modality × tissue × timepoint

Modality	Tissue(s)	Weeks post-FMT
Bacterial metagenomics (long-read)	Fecal (F)	0 (donor), 1, 2, 4, 6, 8
	Colon (C), Ileum (I), Cecum (M)	3, 8
Mycobiome / Virome (long-read)	Fecal (F)	1, 2, 4, 6, 8

This report covers the metagenomics modalities only; metabolomics (LC–MS) and CyTOF immune phenotyping are reported separately.


4. Sample naming convention
Donor sex/ID _ Mouse sex/ID _ Mouse age (weeks) _ Sample type
1. Donor sex/ID: [ F0 | F1 | F2 | F3 | M4 | M5 | F6 | F7 | F8 | F9 | Ctrl (OMM12) | Neg (neg. control)]
2. Mouse sex/ID: [ M | F ] + Mouse ID
3. Mouse age: [ 0 (donor) | 1 | 2 | 3 | 4 | 6 | 8 | X (neg) ]
4. Sample type: [ C (colon) | F (fecal) | I (ileum) | M (cecum) | Ctrl | Neg | 1 (donor) | 2 (donor) ]

Examples:

• F0_F20_6_F: Fecal sample from 6-week-old female mouse #20 derived from female donor #0
• M5_F509_3_M: Cecum sample from 3-week-old female mouse #509 derived from male donor #5
• F6_NA_0_1: 1st fecal sample from female donor #6
• Ctrl_Ctrl3_3_Ctrl: Control sample from the OMM12 group
• Neg_6_X_Neg: Negative control sample

Methods

1. Sequencing and basecalling

   • Long-read sequencing on Oxford Nanopore Technologies (ONT) PromethION flow cells
   • Dorado (v1.1.1): basecalling from POD5 with model dna_r10.4.1_e8.2_400bps_sup@v5.2.0
   • Read filtering policy: reads shorter than 35 bp ignored by Kraken2 during classification; reads longer than 15,000 bp trimmed to 15,000 bp; reads longer than 100,000 bp excluded
   • NanoPlot ^{De Coster and Rademakers (2023)}: per-read length, quality, and yield metrics for long-read data
   • MultiQC ^{Ewels et al. (2016)}: aggregates QC reports from many tools into a single summary report
   • FastQC: per-read quality control checks
   • SAMtools ^{Danecek et al. (2021)}: manipulate and inspect alignment files
2. Donor-derived MAG reference catalogue

   • metaFlye ^{Kolmogorov et al. (2020)}: long-read metagenome assembly using repeat graphs
   • SemiBin2 ^{Pan et al. (2023)}: semi-supervised metagenomic binning
   • Medaka (v2.1.1): consensus polishing of long-read assemblies
   • GTDB-Tk ^{Chaumeil et al. (2022)}: standardized taxonomy assignment for bacterial and archaeal genomes
   • Output: 388 initial bins → 165 non-redundant donor-derived MAGs defining the feature universe for the custom Kraken2 database
3. Taxonomic classifier evaluation and selection

   • Kraken2 ^{Wood et al. (2019)}: k-mer-based per-read taxonomic classification — selected as the primary classifier
   • Minimap2 ^{Li (2018)} + CoverM ^{Aroney et al. (2025)}: alignment-based read mapping and coverage estimation (comparator)
   • MetaPhlAn 4 (Metagenomic Phylogenetic Analysis) ^{Blanco-Míguez et al. (2023)}: marker-gene profiling (comparator)
   • Sylph ^{Shaw and Yu (2025)}: ANI/sketch-based profiling (comparator)
   • Selection criteria: support for fully customizable databases, per-read assignment outputs, and operational scalability under control-sample benchmarking (OMM12 only vs decoy-inclusive databases)
4. Species-level abundance estimation and read-level QC

   • basen (v1.0.0): Base-level Abundance estimation with Species-assigned Evidence using Nanopore — species-rank evidence aggregated from per-read Kraken2 k-mer assignments, normalized by reference genome length to a coverage proxy, and rescaled within sample to relative abundance
   • Read-level compositional QC (basen): two-step filter using (1) the proportion of unassigned k-mers per read (excluded if >75%) and (2) the Shannon diversity of the per-read TaxID assignment profile (threshold >1.0 for high-depth donors, >0.75 for shallower mouse, replicate, and control samples)
5. Mycobiome and virome profiling

   • Reads unclassified against the custom donor-derived Kraken2 database were re-classified against the k2_pluspf (fungi, protozoa) and k2_viral databases
   • Reads unmapped to both fungal and viral indices were screened against the mouse (GRCm39) and human (GRCh38) reference genomes
   • Donor-level cross-domain correlations: Spearman with Phipson–Smyth permutation p-values ^{Phipson and Smyth (2010)}, BH correction within each kingdom-pair family
6. Procrustes analysis of cross-kingdom community structure

   • Symmetric Procrustes superimposition via vegan::procrustes / protest; m² statistic and permutation p-value (999 permutations)
   • Donor-level centroid displacement vectors clustered with Euclidean distance and complete linkage

Results

A. Pre-processing

1. Experimental design
• Oxford Nanopore Technology

Plate	# Barcodesi	# Sequencedii	# Samplesiii
Donors	10	10	10
01	96	96	96
02	96	96	96
03	96	96	96
04	96	96	96
05	96	96	82
06	96	96	95
07	96	96	96
08	96	96	93
09	96	96	92
10	96	96	95
11	96	96	92
12	96	96	96
13	96	96	94
14	41	20	20
Total	1,299	1,278	1,249

Each plate includes one donor, one OMM12 control (barcode96), and one negative control (barcode88), except for the Donors plate.


1. Total number of barcodes sequenced on the PromethION.
2. Twenty one barcodes from Plate 14 yielded zero reads.
   Plate14_barcode01
   Plate14_barcode02
   Plate14_barcode03
   Plate14_barcode04
   Plate14_barcode57
   Plate14_barcode58
   Plate14_barcode59
   Plate14_barcode65
   Plate14_barcode66
   Plate14_barcode67
   Plate14_barcode73
   Plate14_barcode74
   Plate14_barcode75
   Plate14_barcode81
   Plate14_barcode82
   Plate14_barcode83
   Plate14_barcode88
   Plate14_barcode89
   Plate14_barcode90
   Plate14_barcode91
   Plate14_barcode96
3. Twenty nine samples were excluded due to poor quality. These samples showed a read quality value of 0 for all reads in the Dorado summary output.
   Plate05_barcode40
   Plate05_barcode67
   Plate05_barcode85
   Plate05_barcode86
   Plate05_barcode87
   Plate05_barcode88
   Plate05_barcode89
   Plate05_barcode90
   Plate05_barcode91
   Plate05_barcode92
   Plate05_barcode93
   Plate05_barcode94
   Plate05_barcode95
   Plate05_barcode96
   Plate06_barcode88
   Plate08_barcode20
   Plate08_barcode39
   Plate08_barcode50
   Plate09_barcode08
   Plate09_barcode24
   Plate09_barcode32
   Plate09_barcode88
   Plate10_barcode88
   Plate11_barcode24
   Plate11_barcode48
   Plate11_barcode79
   Plate11_barcode80
   Plate13_barcode22
   Plate13_barcode88


2. Basecalling
• MinKNOW does not include the most recent Dorado basecaller.
• Using the POD5 files, we executed the latest release of Dorado (v1.1.1; October 9, 2025) on high-performance computing clusters.
• Model used: dna_r10.4.1_e8.2_400bps_sup@v5.2.0
• A code set for reproducibility: Griffin-Pipeline tutorials
• Workflows
  


3. Basecalling and demultiplex results
• Number of reads in each barcode
  
  


4. Quality assessment
• Dorado Run Analyzer
• Sequencing data summary: reads, quality, N50, etc.
• MultiQC results
• Detailed MultiQC results: (1) per plate, (2) per barcode, and (3) by read length, with reads triaged by length


5. Samples included in the downstream analysis
• Total: 1,249 samples
• 10 donor samples
• 9 donor replicates (low depth)
• 12 OMM12 controls
• 8 negative controls
• 1,210 mice (see Sankey plot below)
  


6. Read quality control
• MultiQC summary: per-base read quality
• Mini conclusion: Based on the MultiQC summary above, we applied length-based read filtering as follows:
1. Short reads were retained, as Kraken2 ignores reads shorter than 35 bp
2. Long reads were trimmed to 15,000 bp
3. Ultra-long reads (>100,000 bp) were discarded


7. Post-QC Stats
• FASTQ stats: number of reads and mean/SD/median quality
• Read length and quality

Number of reads and Quality	Ave. read lenth and Quality

Donor_1 refers to donor samples sequenced on the Donors plate using 10 barcodes of a 96-well set. Donor_2 refers to donor samples sequenced on the other plates (Plate01-14), using one barcode from the same 96-well set.


• Read length and quality per donor

Number of reads	Mean quality	Median quality



• Read length and quality per sample type

Number of reads	Mean quality	Median quality



• Four samples are candidates for exclusion from downstream analysis because of low read counts and/or poor read quality, in addition to eight negative controls.
  F1_F5_3_C
  F7_F707_3_M
  F7_F724_4_F
  F7_F718_8_F
  Neg_1_X_Neg
  Neg_2_X_Neg
  Neg_3_X_Neg
  Neg_4_X_Neg
  Neg_5_X_Neg
  Neg_6_X_Neg
  Neg_7_X_Neg
  Neg_8_X_Neg


8. Classifier evaluation
• Input: control OMM12 samples, N=12 (Ctrl01 - Ctrl12)
• Tools comparison:

Category	Kraken2	Minimap2	MetaPhlAn 4.2	Sylph
Features	k-mer based	alignment-based	marker gene-based	ANI-based
Support long-reads?	yes	yes	yes (v4.2)	yes
Custom DB support?	yes	yes	no	yes
Strengths	high sensitivity/fast classification	high alignment accuracy	low false positives	computationally efficient
Limitations	k-mer ambiguity	computational demands	relies on marker genes	not provide per-read classification
Output type	read-level taxonomic assignment	read-level alignments	taxonomic profile (not per-read classification)	genome-level coverage




Building a custom DB 1: OMM12 (46,004,597 bp) [CP015399] Lachnoclostridium sp. YL32 [CP015400] Ruminiclostridium sp. KB18 [CP015401] Bacteroides sp. I48 [CP015402] Parabacteroides sp. YL27 [CP015403] Burkholderiales bacterium YL45 [CP015404] Erysipelotrichaceae bacterium I46 [CP015405] Blautia sp. YL58 [CP015406] Flavonifractor plautii YL31 [CP015407] Bifidobacterium animalis subsp. animalis YL2 [CP015408] Lactobacillus reuteri I49 [CP015409] Akkermansia muciniphila YL44 [CP015410] Enterococcus faecalis KB1

Building a custom DB 2: DECOY (3,176,927,508 bp) Homo sapiens (GRCh38) Clostridium sp. MD294 strain ASF356 Bacteroides sp. M10 Coprococcus catus GD/7 Bacteroides luhongzhouii Faecalibacterium sp. i21-0019-B1 Roseburia sp. 499 Lacrimispora sp. BS-2 Caproicibacter fermentans Marvinbryantia formatexigens DSM 14469 Chordicoccus furentiruminis Simiaoa sunii strain NSJ-8 OMM12



• Classification results
1. Note 1: MetaPhlAn 4.2, run with its pre-built database, did not classify any reads in the OMM12 samples.
2. Note 2: taxonomic abundance was estimated using CoverM following read alignment with Minimap2.

DB:OMM12	DB:DECOY

The figures above are for the Ctrl01 sample. You can view other samples by clicking the figure.


• Read-level investigation 1: using the OMM12 DB
1. 1:1 comparison of reads assigned to CP015399 (blue) and CP015401 (green) in the Ctrl01 sample
   

   Minimap2+CoverM   CP015399 CP015401 Total Kraken2 CP015399 C1 110,900 C2 2 110,902 CP015401 C3 1,105 C4 464,266 465,371 Total 112,005 464,268 576,273

2. BLAST (Basic Local Alignment Search Tool) was used to examine one of the two reads in cell 2 (C2) that were assigned to CP015399 by Kraken2 but to CP015401 by Minimap2+CoverM
3. Read ID: 0ff4c7d3-1108-4a12-b7ba-5e15912d901e (length: 400 bp / quality: 26.6)

   ATGGAGCCGTAAGCGGCTTTATAGGGATCGTAGGAACCGTCCAGGCCCTTGGCTTTTTTG
   TAGGCCTTAAGCTGCCTGATGATGTGGTTGGCGTATTTGGGATTGTTTTCCGTCACCCAT
   GCCTCAATACCGGAAGTGTCAAAGAGGAGCATGGAAGCTTTTTCCTTATTGATTTTCTGG
   CAGATTGGTTCAGTCAGGTCAACAAGGCGGCTGAACATGGATTGTATGGAAGCAAAGTTC
   ACCTGTGTTGAAATATTCTGTTCCCACGGTTCGATATCCAGTGGATAAGTATCCGTCATC
   ACATCATCCTCATAAAAATACTCGTCGGCCAATCCGCCGAAGCTGTGTCCGAATTCGTGC
   ACTACCACCGGTTTGAACATGGGGTGATGTGCCGTGGTCA

4. BLAST results:
   

   1-226 bp:
   227-400 bp:

5. Minimap2 is unable to align the 1–226 bp segment of the read to CP015399
6. Minimap2 problem definition: Minimap2 relies on minimizers as alignment anchors but discards high-frequency k-mers, and in repetitive regions this anchor loss disrupts chaining and results in missed alignments


• Read-level investigation 2: using the DECOY DB
1. One high-quality read was identified that mapped to the DECOY genome
2. Read ID: 76502669-ef64-431a-97f5-81ae8693f48a (length: 255 bp / quality: 46.2)

   AGCACCTTACTTTATACACATGAGCCATGATATGGGCTTGCATGCGGAAGTGGAAAAGAG
   TCTTCCGGAAAATATTCACCTGGCATTCGACGGATTGGATATCTATCTTTGATTATTGGT
   GAATTTTTTAGTAAAAGTTTTGTTTTCTTTGTAGAAATTATTATTTTTGCACCTAGCCGA
   TAAACAGCGAATATTATGAAACTTCGTACGATAGTGAAAATAGCGATCACGTCTTCTGTT
   GTACTTTTGTGCTCA

3. Kraken2 results:

   Taxon ID	Group	Scientific name	Hit(s)	%
   816		Bacteroides (genus)	71	32.1
   2763022	DECOY	Bacteroides sp. M10	50	22.6
   1796613	OMM12	Bacteroides caecimuris strain I48	49	22.2
   0		Unclassified	51	23.1
   Total			222	100.0

4. Proportion of unclassified reads (erratum: "Reads Group" in the Figure should be replaced with "Database"):
   
5. Shanon diversity index:
   

   Using all taxon IDs assigned to each read	After removing taxon ID 0 (unclassified) from each read



• Mini conclusion: We require a classifier that supports fully customizable databases and avoids biases introduced by sampling or repeat filtering; Kraken2 satisfies both criteria, making it the preferred tool over Minimap2, MetaPhlAn, and Sylph.


9. Create a custom Kraken2 database tailored to the microbial community of interest
• Input: donor samples from the Donor plate (higher depth), N=10 (F0, F2, ..., F9)
• Workflows
  
• Stats:

Donor	# Reads	# Contigs	# Consensus	# Bins	# Passed Bins
F0	4,963,485	13,899	13,899	148	30
F1	3,581,456	7,570	7,570	116	29
F2	4,325,290	11,723	11,723	143	35
F3	7,151,869	18,791	18,791	198	57
M4	8,050,964	11,034	11,034	125	38
M5	6,104,905	9,296	9,296	107	32
F6	4,970,540	11,399	11,399	170	26
F7	3,914,021	11,601	11,601	171	42
F8	2,626,645	11,584	11,584	151	39
F9	3,497,575	14,524	14,521	196	60
Total (distinct)					388	(165)

• Taxonomy labels: GTDB-Tk results and Genome list, N=165


10. Taxonomy classification via Kraken2 with the custom DB (165)
• Input, N=19
• Sequenced using 10 barcodes on a 96-well plate to achieve high(er) coverage (*_NA_0_1), N=10
• An additional set of donor samples was sequenced on separate plates, providing comparable read depth for cross-validation (*_NA_0_2), N=9
F1_NA_0_2 sample was not included in the profiling
• Reference databases:
1. Custom DB with Donors
2. k2_standard: Refseq archaea, bacteria, viral, plasmid, human genome, and a collection of known vectors
3. K2_pluspf: k2_standard plus Refseq protozoa and fungi
• Taxonomy classification:
  
  
  
• Problem definitions:
• Raw taxonomy classification outputs (e.g., read counts) are not directly comparable across samples (sequencing depth varies substantially between runs, flow cells, and barcodes)
• Differences in total classified reads can artificially inflate or deflate apparent taxon abundance (even when the underlying biological composition is similar)
• Taxonomic assignments at different ranks (e.g., genus vs. species) are not on the same scale (preventing consistent comparison across classification levels)


11. Develop a method to calculate relative abundance
• BASEN: Base-level Abundance estimation with Species-assigned Evidence using Nanopore
• Summary: BASEN estimates species-level relative abundance by quantifying species-assigned k-mers from long-read sequencing data, normalizing by reference genome length, and scaling across detected species to obtain genome-size–corrected relative abundances.
• Conceptual figure:
  
  

Species BS GS CS RAS   X 83 100-34 1.2576 0.4666 Y 33 100-34 0.5000 0.1855 Z 15 50-34 0.9375 0.3479 Total 2.7845 1.0000

BS: the total number of species-assigned k-mers for species S GS: total number of bases in the reference genomes of species S CS: approximates the average genome coverage of species S RAS: CS / Σ Ci, where i indexes each species

• Relative abundance: Legend
  
• Correlation among donors:
  


12. Filter low quality reads
• Input: 1,249 samples
• Mice samples, , N=1,210
• Donor samples, N=10
• Donor replicates, N=9
• OMM12 controls, N=12
• Neg. controls, N=8
Relative abundances of the donor replicates and OMM12/negative controls were calculated, but their distributions were omitted from the figures below.
• Filter reads based on unclassified reads, i.e., unaligned k-mers, and Shannon diversity index from Kraken2 k-mer strings (refer to the Results Section 8 above)

Donors	Mice

• Threshold values:
• Unaligned k-mers, proportion: >0.75
• Shannon diversity index: >1.00 (Donor) and >0.75 (Mice, Donor replicates, OMM12 control)

Donors	Mice

B. Microbiome (Figs 1–2)

1. Calculate relative abundance
• All samples in a single heatmap

Capped at 1.0	Capped at 0.5	Capped at 0.1



• Each sample in a single heatmap

Capped	F0	F1	F2	F3	M4	M5	F6	F7	F8	F9
1.0
0.5
0.1



2. Correlation/Variability visualization
• Left: For each factor, all pairwise Spearman correlations were calculated among donor-specific recipient mouse groups, excluding the donor sample. In the radar chart, the yellow and red dots indicate the mean and standard deviation of the correlation coefficients, respectively. The mean ranges from 0.4 to 1.0, while the standard deviation ranges from 0 to 0.15, as indicated in parentheses. For example, cecum samples show higher correlations and greater consistency than ileum samples, as reflected by the lower standard deviation in cecum.
• Right: Violin plots show Spearman correlations between the donor and individual mice based on taxonomic relative abundance. The X-axis indicates the factor, the Y-axis shows the correlation coefficient, and each dot represents the correlation between the donor and a single mouse. Numbers in parentheses denote the number of mice in each factor group.



3. Diversity
• Alpha diversity (within-sample diversity): The Shannon diversity index was calculated for each donor group, sample type, and week group to compare diversity across weeks.
• Per sample visualization
• Across donors and compartments (fecal, colon, cecum, ileum), alpha diversity did not show a consistent time trend: some donor–tissue groups increased at later weeks, while others decreased or remained unchanged, yielding an overall mixed, donor-specific pattern rather than a uniform week effect.

Fecal	Colon
Ileum	Cecum

• Bray–Curtis dissimilarity: Between-sample community dissimilarity based on abundance overlap.
• PCoA (Principal Coordinate Analysis):
    Each point = one sample
    Distances = compositional dissimilarity
    Points closer together = more similar communities
    Points farther apart = more different communities

Fecal	Colon
Ileum	Cecum

• Aitchison distance: Between-sample compositional difference based on log-ratios among taxa (CLR [Centered log ratio] + Euclidean)
• Distance-to-donor distributions by week:
    Smaller = more donor-like
    Larger = more shifted from donor

Type	Distance-to-donor across donors
Fecal
Colon
Ileum
Cecum

• PERMANONA and dispersion:Global and pairwise test results
  
• PCA (Principal Component Analysis)
• Summary of overall compositional variation across samples: PC1 and PC2 capture the largest sources of variation
• Input: CLR (Centered log ratio)-transformed relative abundances: PCA reflects log-ratio balance changes rather than raw proportions
• Note: PCA visualizes variation based on the raw feature space (e.g., transformed taxa abundances), whereas PCoA visualizes differences between samples based on a chosen distance metric (e.g., Bray–Curtis), so PCA reflects variance structure in the data itself, while PCoA reflects ecological dissimilarity between communities.

Fecal	Colon
Ileum	Cecum

• Unexpected patterns from PCA
• Separation along PC1/PC2 suggests a dominant axis of community change; follow-up analyses were performed to determine whether this pattern reflects technical variation or underlying biological factors.

	F1, fecal	F2, fecal
PCA
Aitchison distance
Technical factors
Design (Plate) effect
Mouse sex

• F1 finding: The separation is significantly associated with sequencing depth and read quality, suggesting a technical contribution to the observed clustering.
• F2 finding: No measured technical or metadata factors explain the split, indicating a likely biological or taxon-driven signal and motivating taxon-level follow-up.


4. Engraftment varies over time
• Enumerate a set of particularly prominent taxa
1. The mean relative abundance of each taxon was calculated weekly (e.g., weeks 1, 2, 4, 6, and 8) using fecal samples only, with the donor sample designated as week 0.
2. Using a global cohort-level threshold, taxa with engraftment exceeding 5% of the mean relative abundance were selected.
3. Thirty-one taxa met this criterion in at least one week or donor group and were therefore defined as prominent taxa.

All 123 taxa
Prominent 31 taxa



• Find the most stable engraftment week
1. We calculated Spearman correlation coefficients between donor baseline (week 0) and each post-engraftment week.
2. Then, we selected the week with the highest correlation (most stable engraftment) using the prominent taxa.

All 123 taxa
Prominent 31 taxa



5. Taxa and engraftment characteristics
• Using the donor-specific stable weeks defined below, we compared relative abundances between each donor and the corresponding stable-week recipient profile.
• Week 1: F1
• Week 2: F2, F3, M4
• Week 4: F6, F8
• Week 6: F0, M5, F7, F9
• Week 8: None
• A donor-level threshold of 1% was applied to define engraftment status. Taxa with relative abundance >1% were considered present.
• We annotated taxa as follows:
• ⬤ (closed circle) when both donor and recipient were >1%
• ◯ (open circle) when the taxon was >1% in the donor only
• ✕ (cross) when the taxon was >1% in the recipient only



6. Alpha and beta diversity
• The Shannon diversity index was calculated for each individual sample, encompassing both donor and mouse specimens. The results were subsequently stratified by donor and presented on a weekly basis to illustrate temporal dynamics in alpha diversity across the study period.


• Bray–Curtis dissimilarity was calculated across all samples within the cohort to assess beta diversity.
• The two figures below illustrate distinct clustering patterns based on different grouping variables.

Group by donor	Group by week



7. Convergence
• As the analytical lead, it is necessary to establish a clear and operational definition of convergence based on the dissimilarity matrix.
• Three complementary metrics should be considered:
• Dispersion: Within-week dissimilarity, typically quantified as the distance of each sample to the group centroid.
• Convergence: Between-week dissimilarity, assessing whether samples from consecutive weeks become progressively more similar (i.e., tighter clustering over time).
• Engraftment: Dissimilarity between mouse samples and their corresponding donor samples, reflecting the degree of donor microbiota establishment.
• A single metric is insufficient to adequately define convergence or stability, as each captures distinct structural aspects of the microbial community.



• Convergence can be operationally defined using two primary metrics:
• Dispersion (X-axis): Distance to centroid, representing within-group variability.
• Convergence (Y-axis): Dissimilarity between an anchor week and subsequent week samples, representing temporal tightening or alignment.

Group by donor	Group by week

• A two-dimensional plot below incorporating engraftment (distance to donor) on the X-axis and cohesion on the Y-axis can provide informative comparisons across donors.
• However, engraftment represents a distinct biological construct—specifically, the degree of similarity to the donor microbiota—rather than the internal stabilization dynamics of the microbial community. Because engraftment reflects donor alignment rather than structural convergence within the cohort, it conveys a different conceptual meaning.
• Therefore, in this study, convergence is defined solely on the basis of "Within-week dispersion" and "Between-week convergence", which together capture intra-group variability and temporal tightening of community structure.

Discussion

• Donor-dominated variation. Donor identity is the dominant axis of variation in the recipient gut microbiome and its functional potential.
• Engraftment long tail. 42 of 165 donor MAGs absent from recipients; 92 of 123 engrafted MAGs stayed below the 5% prominence threshold throughout follow-up.
• Pipeline contribution. Donor-tailored Kraken2 reference + basen genome-size-aware abundance estimation resolves the low-abundance long tail and is portable to any project-specific reference setting.
• Microbiome structure. Donor-partitioned fecal communities; alpha-diversity ranks preserved across Colon/Ileum/Cecum; functional Spearman > 0.96 to donor stool at each donor’s modal stable week; Ileum least represented by fecal sampling; bacterial × viral structure > bacterial × fungal at this depth.

Limitations

• Only one donor technical replicate at comparable depth/quality — limits variance partitioning.
• Fecal microbiome series anchored to Weeks 1, 2, 4, 6, 8; no fecal sampling at Week 3 (terminal Colon/Ileum/Cecum collected at Week 3 instead).
• Minimal filtering preserves exploratory variability; downstream inferential reuse requires question-specific thresholds.
• Mycobiome (0.02%) and virome (0.17%) reads are minor at this sequencing depth — cross-kingdom power is correspondingly limited.
• Feature universe is donor-bounded: taxa below assembly detection in donor stool are not in the recipient classification space.

Data Availability

• [SRA (Sequence Read Archive)] pending
• [Shiny app] MICROBENCH

Code Availability

• [Zenodo] Scripts
• [GitHub] Griffin-Pipeline
• [GitHub] basen (Base-level Abundance estimation with Species-assigned Evidence using Nanopore)

References

Aroney et al. CoverM: read alignment statistics for metagenomics. Bioinformatics 2025;41(4):btaf147.
Blanco-Míguez et al. Extending and improving metagenomic taxonomic profiling with uncharacterized species using MetaPhlAn 4. Nat Biotechnol 2023;41(11):1633-1644.
Chaumeil et al. GTDB-Tk v2: memory friendly classification with the genome taxonomy database. Bioinformatics 2022;38(23):5315-5316.
Danecek et al. Twelve years of SAMtools and BCFtools. Gigascience 2021;10(2):giab008.
De Coster and Rademakers. NanoPack2: population-scale evaluation of long-read sequencing data. Bioinformatics 2023;39(5):btad311.
Ewels et al. MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics 2016;32(19):3047-8.
Kolmogorov et al. metaFlye: scalable long-read metagenome assembly using repeat graphs. Nat Methods 2020;17(11):1103-1110.
Li. Minimap2: pairwise alignment for nucleotide sequences. Bioinformatics 2018;34(18):3094-3100.
Pan et al. A deep siamese neural network improves metagenome-assembled genomes in microbiome datasets across different environments. Nat Commun 2022;13(1):2326.
Phipson and Smyth. Permutation P-values should never be zero: calculating exact P-values when permutations are randomly drawn. Stat Appl Genet Mol Biol 2010;9(1):Article 39.
Seo et al. Dual platform spatial transcriptomics reveals parvalbumin interneuron subtype vulnerability in mouse models of Alzheimer’s disease. Nat Commun 2026.
Shaw and Yu. Rapid species-level metagenome profiling and containment estimation with sylph. Nat Biotechnol 2025;43(8):1348-1359.
Walter et al. Establishing or exaggerating causality for the gut microbiome: Lessons from human microbiota-associated rodents. Cell 2020;180(2):221-232.
Wood et al. Improved metagenomic analysis with Kraken 2. Genome Biol 2019;20(1):257.

Abbreviations