Dataset DOI: 10.5061/dryad.gtht76j0x

Description of the data and file structure

Data were collected to characterize how deletion of phosphoglycerate mutase 5 alters the lipidome and/or lipid metabolism related transcriptome in human hepatocellular adenoma (HepG2, scRNAseq only) / carcinoma (Huh7, both lipidomics and scRNAseq) cell lines.

Files and variables

File name: lipidomics_data.zip

Overview_ Lipidomics Data (Huh7 cells)

This dataset contains lipidomics profiling data generated from Huh7 cell samples treated with bovine serum albumin (BSA). Both positive and negative ionization mode datasets are included, along with combined tables, statistical analyses, and pathway results. Supporting files provide details of sample preparation and analysis summaries. 

Units and measurement conventions

• Neutral mass (Da): daltons
• m/z: mass-to-charge ratio; reported numerically and commonly treated as unitless
• Charge: integer charge state, unitless
• Retention time (min): minutes
• Drift time (ms): milliseconds
• CCS (angstrom^2): collision cross section in square angstroms
• dCCS (angstrom^2): CCS difference or error term in square angstroms
• Chromatographic peak width (min): minutes
• Anova (p) and q Value: unitless probabilities
• Max Fold Change: unitless ratio
• Minimum CV%: percent
• Score, Fragmentation Score, Isotope Similarity: unitless software-derived scores
• Mass Error (ppm): parts per million
• Retention Time Error (mins): minutes
• Normalized abundance, Raw abundance, and Maximum Abundance: relative ion abundance in arbitrary units (a.u.)

Raw/Processed Compound Lists

These are standard outputs.

34204_Lipidomics_Negative_Compounds.csv 

  Feature table of identified/quantified lipids in negative ion mode. Columns include:

• m/z (mass-to-charge ratio) - unitless (Da/z)
• Retention time (RT) - minutes (min)
• Peak intensity / abundance - arbitrary units (a.u., ion counts)
• Adduct type
• Compound name / lipid class

34204_Lipidomics_Positive_Compounds.csv 

  Feature table of identified/quantified lipids in positive ion mode. Columns include:

• m/z (mass-to-charge ratio) - unitless (Da/z)
• Retention time (RT) - minutes (min)
• Peak intensity / abundance - arbitrary units (a.u., ion counts)
• Adduct type
• Compound name / lipid class

Combined Input-Negative and Positive (1).csv

  Merged dataset combining both ionization modes.

Analysis Outputs

EzInfo statistics data.xlsx

  Multivariate statistical analysis results (e.g., PCA, PLS-DA) generated with EzInfo software. Includes:

• p-values 
• Fold change (FC) 
• p(corr) - correlation-scaled loading
• VIP scores

S-plot thresholds:

• Features contributing to group separation defined by:
  • p(corr) > 0.5
  • p(corr) < -0.5

MetaboAnalyst_result_pathway.csv

  Pathway enrichment/impact analysis results exported from MetaboAnalyst. Contains pathway enrichment results:

• Pathway impact score 

• p-values (enrichment significance) 

• -log(p) 

  
  

Summary with statistics.pptx

Summary of statistical findings and key lipidomic changes. Includes:

• PCA score plots (Hotelling’s T² ellipse)
• S-plots highlighting discriminating lipids
• Separation between experimental groups in both ion modes 
• Data (Parameter- unit)
  • m/z- Da/z
  • Retention time-minutes (min)
  • Peak intensity arbitrary units (a.u.)
  • Injection volume- µL
  • Solvent volumes- µL
  • Collision energy- Volts (V)
  • Capillary voltage- kilovolts (kV)
  • Concentration additives- mM, % (v/v)
  • p-values- unitless
  • Fold change- unitless
  • Pathway impact- unitless

Supporting Information

Sample preparation.pdf

  Description of experimental setup, cell culture, extraction protocols, and instrument settings.

• .DS_Store (system file, can be ignored)

 

Data Usage Notes

• Raw compound tables contain m/z, retention time, peak area/intensity, and annotation columns. 
• Combined dataset may serve as input for downstream multi-omics integration or pathway enrichment. 
• Statistical analysis files provide ready-to-use outputs but should be cross-checked with raw data for reproducibility. 

 

---

 

Recommended Workflow

1. Start with the Sample preparation.pdf for experimental context. 
2. Use Negative and Positive Compounds CSVs as the primary raw data. 
3. Refer to Combined Input CSV for integrative analysis. 
4. Review EzInfo and MetaboAnalyst outputs for statistical/pathway insights. 
5. Consult the Summary PDF for an overview of results. 

 

---

Metadata

• Organism/Cell line: Huh7 (human hepatoma cells) 
• Technique: LC-MS/MS lipidomics (positive & negative ion modes) 
• Data formats: CSV, XLSX, PDF 

Date: November 2024

Contact: john8926@msu.edu

File: scRNAseq_data.zip

Folders

In the "HepG2 scRNA data" and the "Huh7 scRNA data" folders:

The single cell RNA (scRNA) sequencing data for the HepG2 (hepatocellular adenoma cell line) or Huh7 (hepatocellular carcinoma) and LX2 (hepatic stellate cell line) samples.

In the "filter_housekeeping" and "filter_housekeeping_and_MT" (MT, mitochondria) folders:

Pre-processed data devoid of housekeeping genes or housekeeping genes and mitochondrial genes are included. Each sub-folder is divided into the "cell_cycle_correction" and "no_correction" folders which include the data with and without cell cycle correction.

In the "cell_cycle_correction" and "no_correction" folders:

Each file is subdivided into key data features including cell cluster highlights, markers of each cluster, and summaries of differentially expressed genes per treatment and genotype.

In the "cluster_highlight_plots" folder:

This dataset contains a series of UMAP visualizations representing single-cell transcriptomic data from cell line samples. Each file highlights the distribution of cells from a specific sample overlaid on a shared embedding. The figures illustrate how individual datasets map onto a common low-dimensional space, enabling comparison of cell populations (clusters) across cell types and cell cycle phase.

In the "cluster_markers/" folder:

all_cluster_markers_custom_typing.csv- This file contains the results of the FindAllMarkers function in Seurat, used to find markers (differentially expressed genes) for each of the cell clusters identified by Seurat. We have added our cell typing results. It contains 9 columns: "cluster" is the cell cluster, as identified by Seurat. "types_split"- is the cell type, but maintaining the sub-clustering found by Seurat. "types_grouped" is the cell type, lumping all clusters of the overall cell type together. This column is just for ease of idenfication and sorting: all statistical results in the file were performed at the level of the split clusters. "avg_log2FC" is the average difference in log2-transformed expression between this cluster and all other clusters: a positive value means the gene is more highly expressed in the cluster, negative values mean it is less expressed within the cluster. "pct.1" is the percentage of cells in which the gene is detected in the cluster, and "pct.2" is the percentage of cells in which the gene is detected in all other clusters. "p_val" is the raw p-value from the statistical test for differential expression (here, a Wilcoxon Rank Sum test), and "p_val_adj" is the Bonferroni-corrected p-value. Bonferroni correction was done within each cluster (not at the experiment-wide level).

top20_per_cluster_custom_typing.csv- This file also contains output from the FindAllMarkers function, but is filtered down to the top 20 upregulated genes for each cluster. That is, the top 20 marker candidates for each cluster. The columns are the same as in all_cluster_markers_custom_typing.csv.

In the "DE_within_cell_types" folder:

Each file within the folder gives results from one of the 4 desired differential expression analyses: looking at differential expression between genotypes in the BSA treatment, looking at DE between genotypes in the palmitate (PA) treatment, looking at DE between treatments in the wild-type (WT) genotype, and looking at DE between treatments in the knockout (KO) genotype. These differential expression analyses are performed separately for each identified cell type. So, there are 4 .csv files per cell type.

These .csv files contain 10 columns. "cell_type" gives the cell type within which the statistical test is being preformed. "factor" gives the statistical factor being compared. "group_1" and "group_2" give the two groups that are being compared for differential expression. "gene" is the gene being tested for differential expression. "avg_log2FC" is the average difference in log2-transformed expression between the two groups: a positive value means the gene is more highly expressed in group 1, negative values mean it is more highly expressed in group 2. "pct.1" is the percentage of cells in which the gene is detected in group 1, and "pct.2" is the same for group 2. "p_val" is the raw p-value from the statistical test for differential expression (here, a Wilcoxon Rank Sum test), and "p_val_adj" is the Bonferroni-corrected p-value. Bonferroni correction was done within each cell type by factor combination (not at the experiment-wide level).

Metadata

• Organism/Cell line: Huh7 (human hepatocellular carcinoma cells), HepG2 (human hepatocellular adenoma cells), LX2 (human hepatic stellate cells)
• Technique: single cell RNAseq on hepatocellular adenoma/carcinoma cells co-cultured with mesenchymal cells and treated with palmitate conjugated to bovine serum albumin (PA-BSA) or BSA (control). Only the Huh7 control cells are used for lipidome correlation.
• Data formats: CSV, XLSX, PDF 

Dates: May 2022 (Huh7), June 2022 (HepG2)

 Contact: john8926@msu.edu

Code/software

The file data formats include CSV, XLSX, PDF 

Files can be opened with Microsoft and Adobe software, or multiple online data viewers.

Access information

N/A