1 Executive Summary

This report presents untargeted LC-MS metabolomics analysis of the supernatant fraction of Bacillus amyloliquefaciens BNC5 cultures, comparing three conditions: BNC5 alone, BNC5 co-cultured with Ralstonia solanacearum whole cells (BNC5RSw), and RSw alone. The supernatant fraction captures secreted and extracellular metabolites, complementing the intracellular metabolome characterised in the parallel pellet fraction analysis (rpubs.com/Onuh007/1453638).

Key findings:

  • 5235 features showed significant variation across groups (ANOVA, BH-adjusted p < 0.05)
  • 560 features were elevated in BNC5RSw vs BNC5 (log2FC > 0.58, p_BH < 0.05)
  • After removing 176 RSw background features, 384 true BNC5 secreted response features were identified
  • 277 features were significantly reduced in BNC5RSw vs BNC5, indicating altered secretion patterns

2 Study Design

Table 1: Experimental design — supernatant fraction
Condition Description Fraction Replicates
BNC5 B. amyloliquefaciens BNC5 alone — negative control Supernatant 6 (Lot1 x3 + Lot2 x3)
BNC5RSw BNC5 co-cultured with R. solanacearum whole cells — treatment Supernatant 6 (Lot1 x3 + Lot2 x3)
RSw R. solanacearum whole cells alone — background control Supernatant 6 (Lot1 x3 + Lot2 x3)

Sample type: Bacterial culture supernatant (extracellular fraction)
Instrument: HPLC-QTOF ESI-MS (positive ionisation mode)
Biological lots: 2 independent preparations per condition
Total samples analysed: 18

3 Data Processing

3.1 XCMS Peak Detection

Raw mzXML files were processed using XCMS (v4.10.1):

  • Peak detection: CentWave algorithm
  • Retention time correction: LOESS smoothing, span = 0.6, 2 rounds
  • Peak grouping: density method, bandwidth = 10
  • Missing peak filling: fillPeaks()
  • Adduct and isotope annotation: CAMERA (positive mode)
  • Total features detected: 10199

3.2 Preprocessing Pipeline

Table 2: Step-by-step preprocessing summary
Step Result
Raw XCMS features 10199 features detected
Zeros converted to NA 8628 zeros replaced with NA
50% per-group presence filter 10199 features retained — 0 removed
Half-minimum imputation All missing values replaced — 0 NAs remaining
TIC normalisation All 18 samples normalised to equal total ion count
Log2 transformation Range: -28.4 to -4.5
Autoscaling (z-score per feature) Mean = 0, SD = 1 per feature across samples

3.3 Normalisation Quality Check

Figure 1: Sample-level intensity distributions before and after preprocessing. After TIC normalisation and log2 transformation, samples show comparable distributions. Autoscaling centres all features at zero.

Figure 1: Sample-level intensity distributions before and after preprocessing. After TIC normalisation and log2 transformation, samples show comparable distributions. Autoscaling centres all features at zero.

4 Quality Control

4.1 Lot Effect Assessment

Figure 2: PCA coloured by analytical lot. Overlap between Lot1 and Lot2 confirms no significant technical batch effect between the two independent biological replicates.

Figure 2: PCA coloured by analytical lot. Overlap between Lot1 and Lot2 confirms no significant technical batch effect between the two independent biological replicates.

Lot1 and Lot2 samples overlapped completely in PCA space confirming no significant technical effect between the two independent biological lots. Lot was not included as a covariate in statistical models.

4.2 Biological Group Separation

Figure 3: PCA scores plot coloured by biological condition. All three groups show clear separation confirming distinct metabolic profiles between conditions.

Figure 3: PCA scores plot coloured by biological condition. All three groups show clear separation confirming distinct metabolic profiles between conditions.

PC1 explained 39.3% of total variance and clearly separated the three biological conditions. PC2 explained a further 18.7%, together accounting for 58% of total variance.

5 Statistical Analysis

5.1 ANOVA Results

5235 of 10199 features showed significant variation across the three conditions after Benjamini-Hochberg correction (p < 0.05).

Table 3: Top 15 features by ANOVA significance (BH-corrected)
Feature Raw p-value BH-adjusted p Significant
M363.2320T1396.82 4.21e-20 4.30e-16 TRUE
M417.2622T1295.90 3.73e-19 1.90e-15 TRUE
M420.2629T1308.50 1.90e-18 6.46e-15 TRUE
M1037.4274T2462.29 2.98e-18 7.61e-15 TRUE
M179.1783T1400.01 4.21e-18 8.60e-15 TRUE
M1572.0643T2324.85 9.95e-18 1.69e-14 TRUE
M439.2470T1395.26 1.38e-17 2.01e-14 TRUE
M335.2225T1526.08 1.79e-17 2.28e-14 TRUE
M225.1332T1309.78 3.08e-17 3.37e-14 TRUE
M265.1805T1714.36 3.64e-17 3.37e-14 TRUE
M906.5088T1740.27 3.59e-17 3.37e-14 TRUE
M566.3511T1386.53 4.26e-17 3.62e-14 TRUE
M311.2598T1785.56 6.33e-17 4.97e-14 TRUE
M235.1176T1310.15 8.88e-17 6.46e-14 TRUE
M1466.0532T2525.70 9.50e-17 6.46e-14 TRUE

5.2 RSw Background Subtraction

To identify metabolites specifically secreted by BNC5 in response to RSw challenge, a three-step filter was applied. This is essential in supernatant analysis because the extracellular medium in the BNC5RSw condition contains both BNC5-secreted compounds AND metabolites leached from the RSw dead cells:

Filter Step Count
Elevated in BNC5RSw vs BNC5 (log2FC > 0.58, p_BH < 0.05) 560
Removed — also elevated in RSw background 176
True BNC5 secreted response features 384
Features reduced in BNC5RSw vs BNC5 277

5.3 Volcano Plot

Figure 4: Volcano plot of BNC5RSw vs BNC5 supernatant. Blue = true BNC5 secreted response features after RSw background removal. Orange = RSw background features excluded from analysis. Red = features reduced in BNC5RSw. Grey = not significant.

Figure 4: Volcano plot of BNC5RSw vs BNC5 supernatant. Blue = true BNC5 secreted response features after RSw background removal. Orange = RSw background features excluded from analysis. Red = features reduced in BNC5RSw. Grey = not significant.

6 Multivariate Analysis

6.1 PLS-DA Scores Plot

Figure 5: PLS-DA scores plot using the true BNC5 secreted response features only. Complete separation of all three groups confirms the biological specificity of the identified secreted metabolome.

Figure 5: PLS-DA scores plot using the true BNC5 secreted response features only. Complete separation of all three groups confirms the biological specificity of the identified secreted metabolome.

6.2 VIP Scores

Figure 6: PLS-DA VIP scores for the top 30 secreted BNC5 response features. Features with VIP > 1.0 (blue bars) contribute above average to group discrimination.

Figure 6: PLS-DA VIP scores for the top 30 secreted BNC5 response features. Features with VIP > 1.0 (blue bars) contribute above average to group discrimination.

240 features showed VIP scores above 1.0, indicating above-average contribution to discrimination between biological conditions.

7 Top Secreted BNC5 Response Features

Table 4: Top 20 true BNC5 secreted response features — ranked by VIP score
Feature m/z RT (min) Log2FC Fold Change p_BH VIP Score
M1143.8958T2521.58 1143.8958 42.03 5.791 55.36 4.91e-06 1.081
M564.5051T2539.45 564.5051 42.32 3.657 12.61 5.23e-05 1.081
M578.5182T2564.97 578.5182 42.75 2.436 5.41 3.46e-05 1.079
M1433.0475T2522.46 1433.0475 42.04 8.257 305.98 4.17e-05 1.079
M714.5000T2498.09 714.5000 41.63 2.446 5.45 4.60e-05 1.079
M733.5154T2450.82 733.5154 40.85 2.922 7.58 1.41e-06 1.078
M1419.0290T2537.34 1419.0290 42.29 4.643 24.98 1.35e-05 1.078
M1411.0490T2537.57 1411.0490 42.29 4.848 28.81 3.30e-05 1.078
M1422.0524T2535.50 1422.0524 42.26 6.877 117.55 6.99e-05 1.078
M576.5057T2525.58 576.5057 42.09 3.883 14.75 1.79e-04 1.078
M1037.4274T2462.29 1037.4274 41.04 7.601 194.21 1.41e-06 1.077
M1131.8927T2536.68 1131.8927 42.28 3.747 13.42 9.58e-06 1.077
M1430.0245T2536.51 1430.0245 42.28 4.700 25.99 5.37e-05 1.077
M688.4911T2457.67 688.4911 40.96 3.656 12.61 6.51e-05 1.077
M619.1305T1621.16 619.1305 27.02 6.432 86.36 1.02e-04 1.077
M703.5094T2493.16 703.5094 41.55 4.217 18.59 1.65e-04 1.077
M591.1906T1591.89 591.1906 26.53 5.813 56.22 2.89e-07 1.076
M755.5525T2598.19 755.5525 43.30 2.646 6.26 8.13e-07 1.076
M689.4938T2457.89 689.4938 40.96 3.720 13.18 1.02e-04 1.076
M1460.0689T2555.47 1460.0689 42.59 9.116 554.86 1.51e-04 1.076

8 Pellet vs Supernatant Comparison

This supernatant analysis complements the parallel pellet fraction analysis (rpubs.com/Onuh007/1453638). Together they provide a complete picture of BNC5 metabolic response to RSw:

Fraction What it measures Features identified
Pellet Intracellular metabolic reprogramming 384 (this fraction)
Supernatant Secreted defensive metabolites See pellet report

Key biological questions from integrated analysis:

  • Features elevated in both fractions — produced intracellularly AND actively secreted
  • Features elevated in pellet only — intracellular accumulation without secretion
  • Features elevated in supernatant only — direct secretion without intracellular accumulation
  • This cross-fraction comparison will identify the primary secreted defensive compounds

9 Conclusions

Untargeted LC-MS metabolomics of the supernatant fraction of B. amyloliquefaciens BNC5 identified 384 features specifically elevated upon exposure to R. solanacearum whole cells, after rigorous removal of RSw background metabolites (n = 176).

A further 277 features were significantly reduced in BNC5RSw supernatant compared to BNC5 alone, suggesting that BNC5 alters its secretion profile during pathogen interaction — possibly diverting resources from normal secretory metabolism toward production of defensive compounds.

The complete absence of lot effect confirmed biological reproducibility across two independent experimental preparations. PCA and PLS-DA both confirmed complete metabolic separation between all three conditions, validating the experimental design and analytical workflow.

Recommended next steps:

  1. Putative annotation of 384 secreted response features using accurate mass database searching (HMDB, METLIN, LipidMaps) and MS/MS fragmentation confirmation for top candidates
  2. Integrated pellet-supernatant analysis to identify which compounds are produced intracellularly and secreted versus retained
  3. Negative ionisation mode analysis for complementary coverage of lipids and organic acids
  4. Targeted quantitative LC-MS/MS for validation of top candidate compounds

Analysis performed by Augustine Onuh, PhD
Department of Chemistry, Chulalongkorn University, Bangkok, Thailand
Generated: 2026-09-23 | R version 4.6.0 | xcms v4.10.1
Pellet report: rpubs.com/Onuh007/1453638
GitHub: github.com/Onuh007