# Characterizing Human Cell Types and Tissue Origin Using the Benford Law

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Materials and Methods

#### 2.1. Datasets

#### 2.2. Analysis of Benford Distribution

#### 2.3. Lists of Genes

#### 2.4. Statistical Analysis

#### 2.5. Clustering and Machine Learning

## 3. Results

## 4. Discussion

## Supplementary Materials

## Author Contributions

## Funding

## Conflicts of Interest

## References

- Newcomb, S. Note on the Frequency of Use of the Different Digits in Natural Numbers. Am. J. Math.
**1881**, 4, 39–40. [Google Scholar] [CrossRef] - Benford, F. The Law of Anomalous Numbers. Proc. Am. Philos. Soc.
**1938**, 78, 551–572. [Google Scholar] - Nigrini, M.J. I’ve got your number. J. Account.
**1999**, 187, 79–83. [Google Scholar] - Nigrini, M.J. Benford’s Law: Applications for Forensic Accounting, Auditing, and Fraud Detection; Wiley: Hoboken, NJ, USA, 2012. [Google Scholar]
- Kreuzer, M.; Jordan, D.; Antkowiak, B.; Drexler, B.; Kochs, E.F.; Schneider, G. Brain Electrical Activity Obeys Benford’s Law. Anesth. Analg.
**2014**, 118, 183–191. [Google Scholar] [CrossRef] [PubMed] - Friar, J.L.; Goldman, T.; Pérez–Mercader, J. Genome Sizes and the Benford Distribution. PLoS ONE
**2012**. [Google Scholar] [CrossRef] - Hoyle, D.C.; Rattray, M.; Jupp, R.; Brass, A. Making sense of microarray data distributions. Bioinformatics
**2002**, 18, 576–584. [Google Scholar] [CrossRef] - Sandron, F.; Hayford, S.R. Do Populations Conform to the Law of Anomalous Numbers? Popululation
**2002**, 57, 753–761. [Google Scholar] [CrossRef] - Costas, E.; López-Rodasa, V.; Toro, J.F.; Flores-Moya, A. The number of cells in colonies of the cyanobacterium Microcystis aeruginosa satisfies Benford’s law. Aquat. Bot.
**2008**, 89, 341–343. [Google Scholar] [CrossRef] - Whyman, G.; Shulzinger, E.; Bormashenko, E. Intuitive considerations clarifying the origin and applicability of the Benford law. Results Phys.
**2016**, 6, 3–6. [Google Scholar] [CrossRef][Green Version] - Pericchi, L.; Torres, D.; Student, P.D. Quick Anomaly Detection by the Newcomb-Benford Law, with Applications to Electoral Processes Data from the USA, Puerto Rico and Venezuela. Stat. Sci.
**2011**, 26, 502–516. [Google Scholar] [CrossRef] - Cerioli, A.; Barabesi, L.; Cerasa, A.; Menegatti, M.; Perrotta, D. Newcomb–Benford law and the detection of frauds in international trade. Proc. Natl. Acad. Sci. USA
**2019**, 116, 106–115. [Google Scholar] [CrossRef] [PubMed] - Karthik, D.; Stelzer, G.; Gershanov, S.; Baranes, D.; Salmon-Divon, M. Elucidating tissue specific genes using the Benford distribution. BMC Genom.
**2016**. [Google Scholar] [CrossRef] - Lun, A.T.L.; McCarthy, D.J.; Marioni, J.C. A step-by-step workflow for low-level analysis of single-cell RNA-seq data with Bioconductor. F1000Research
**2016**. [Google Scholar] [CrossRef] - Stuart, T.; Butler, A.; Hoffman, P.; Hafemeister, C.; Papalexi, E.; Mauck, W.M.; Hao, Y.; Stoeckius, M.; Smibert, P.; Satija, R. Comprehensive Integration of Single-Cell Data. Cell
**2019**, 177, 1888–1902. [Google Scholar] [CrossRef] [PubMed] - Wagner, F.; Yanai, I. Moana: A robust and scalable cell type classification framework for single-cell RNA-Seq data. bioRxiv
**2018**. [Google Scholar] [CrossRef] - Alavi, A.; Ruffalo, M.; Parvangada, A.; Huang, Z.; Bar-Joseph, Z. A web server for comparative analysis of single-cell RNA-seq data. Nat. Commun.
**2018**. [Google Scholar] [CrossRef] [PubMed] - Pollen, A.A.; Nowakowski, T.J.; Shuga, J.; Wang, X.; Leyrat, A.A.; Lui, J.H.; Li, N.; Szpankowski, L.; Fowler, B.; Chen, P.; et al. Low-coverage single-cell mRNA sequencing reveals cellular heterogeneity and activated signaling pathways in developing cerebral cortex. Nat. Biotechnol.
**2014**, 32, 1053–1058. [Google Scholar] [CrossRef] [PubMed][Green Version] - Rizzetto, S.; Eltahla, A.A.; Lin, P.; Bull, R.; Lloyd, A.R.; Ho, J.W.K.; Venturi, V.; Luciani, F. Impact of sequencing depth and read length on single cell RNA sequencing data of T cells. Sci. Rep.
**2017**. [Google Scholar] [CrossRef] [PubMed] - Chu, L.F.; Leng, N.; Zhang, J.; Hou, Z.; Mamott, D.; Vereide, D.T.; Choi, J.; Kendziorski, C.; Stewart, R.; Thomson, J.A. Single-cell RNA-seq reveals novel regulators of human embryonic stem cell differentiation to definitive endoderm. Genome Biol.
**2016**. [Google Scholar] [CrossRef] - Dobin, A.; Davis, C.A.; Schlesinger, F.; Drenkow, J.; Zaleski, C.; Jha, S.; Batut, P.; Chaisson, M.; Gingeras, T.R. STAR: Ultrafast universal RNA-seq aligner. Bioinformatics
**2013**, 29, 15–21. [Google Scholar] [CrossRef] - Liao, Y.; Smyth, G.K.; Shi, W. featureCounts: An efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics
**2014**, 30, 923–930. [Google Scholar] [CrossRef] [PubMed] - GTEx Portal. Available online: https://gtexportal.org/home/ (accessed on 23 August 2019).
- R Development Core Team R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2018.
- Joenssen, D.W. Statistical Tests for Evaluating Conformity to Benford’s Law. Available online: https://rdrr.io/cran/BenfordTests/ (accessed on 27 August 2019).
- Palmer, N.P.; Schmid, P.R.; Berger, B.; Kohane, I.S. A gene expression profile of stem cell pluripotentiality and differentiation is conserved across diverse solid and hematopoietic cancers. Genome Biol.
**2012**. [Google Scholar] [CrossRef] [PubMed] - Satija, R.; Farrell, J.A.; Gennert, D.; Schier, A.F.; Regev, A. Spatial reconstruction of single-cell gene expression data. Nat. Biotechnol.
**2015**, 33, 495–502. [Google Scholar] [CrossRef] [PubMed] - Finak, G.; McDavid, A.; Yajima, M.; Deng, J.; Gersuk, V.; Shalek, A.K.; Slichter, C.K.; Miller, H.W.; McElrath, M.J.; Prlic, M.; et al. MAST: A flexible statistical framework for assessing transcriptional changes and characterizing heterogeneity in single-cell RNA sequencing data. Genome Biol.
**2015**. [Google Scholar] [CrossRef] [PubMed] - Sonnenblick, A.; Brohée, S.; Fumagalli, D.; Vincent, D.; Venet, D.; Ignatiadis, M.; Salgado, R.; Van den Eynden, G.; Rothé, F.; Desmedt, C.; et al. Constitutive phosphorylated STAT3-associated gene signature is predictive for trastuzumab resistance in primary HER2-positive breast cancer. BMC Med.
**2015**. [Google Scholar] [CrossRef] [PubMed] - Van Der Maaten, L.; Hinton, G. Visualizing Data using t-SNE. J. Mach. Learn. Res.
**2008**, 9, 2579–2605. [Google Scholar] - Kuhn, M. Building Predictive Models in R Using the caret Package. J. Stat. Softw.
**2008**, 28, 1–26. [Google Scholar] [CrossRef] - Robin, X.; Turck, N.; Hainard, A.; Tiberti, N.; Lisacek, F.; Sanchez, J.-C.; Müller, M. pROC: An open-source package for R and S+ to analyze and compare ROC curves. BMC Bioinform.
**2011**. [Google Scholar] [CrossRef] - Kafri, O. Cornell (Unversity Ithaca, NY, USA) Entropy Principle in Direct Derivation of Benford’s Law.
**2009**. Unpublished work. [Google Scholar] - Shekhar, K.; Lapan, S.W.; Whitney, I.E.; Tran, N.M.; Macosko, E.Z.; Kowalczyk, M.; Adiconis, X.; Levin, J.Z.; Nemesh, J.; Goldman, M.; et al. Comprehensive Classification of Retinal Bipolar Neurons by Single-Cell Transcriptomics. Cell
**2016**, 166, 1308–1323. [Google Scholar] [CrossRef] - Chen, H.-I.H.; Jin, Y.; Huang, Y.; Chen, Y. Detection of high variability in gene expression from single-cell RNA-seq profiling. BMC Genomics
**2016**. [Google Scholar] [CrossRef] [PubMed] - Usoskin, D.; Furlan, A.; Islam, S.; Abdo, H.; Lönnerberg, P.; Lou, D.; Hjerling-Leffler, J.; Haeggström, J.; Kharchenko, O.; Kharchenko, P.V.; et al. Unbiased classification of sensory neuron types by large-scale single-cell RNA sequencing. Nat. Neurosci.
**2015**, 18, 145–153. [Google Scholar] [CrossRef] - Wu, B. Differential gene expression detection and sample classification using penalized linear regression models. Bioinformatics
**2006**, 22, 472–476. [Google Scholar] [CrossRef] [PubMed] - Luecken, M.D.; Theis, F.J. Current best practices in single-cell RNA-seq analysis: A tutorial. Mol. Syst. Biol.
**2019**. [Google Scholar] [CrossRef] [PubMed]

**Figure 1.**Correlations between scRNA-seq data and the Benford distribution. (

**a**) Predicted prevalence of the occurrence of each leading digit according to the Benford Law. D = digit; P(Digit) = Frequency of D. (

**b**) Gene-centered and (

**c**) cell-centered first-digit distributions, calculated across 1,018 single cells. The predicted Benford distribution is represented by red dots. (

**d**) First-digit distribution of gene-centered expression data of six cell types. White circles represent the predicted Benford distribution. H1: human embryonic stem cells (n = 375 cells); NPC: neural progenitor cells (n = 173 cells); DEP: definitive endoderm progenitors (n = 138 cells); EC: endothelial cells (n = 105 cells); TB: trophoblasts (n = 69 cells); HFF: human foreskin fibroblasts (n = 159 cells).

**Figure 2.**Distribution of group-based, cell-centered (

**a**) MAE and (

**b**) mean-expression scores, calculated across all genes (left) and across 178 pluripotent genes (right) [26]. H1: human embryonic stem cells (n = 375 cells); NPC: neural progenitor cells (n = 173 cells); DEP: definitive endoderm progenitors (n = 138 cells); EC: endothelial cells (n = 105 cells); TB: trophoblasts (n = 69 cells); HFF: human foreskin fibroblasts (n = 159 cells). Each group was compared to all other cell types, combined, using a Wilcoxon test. Ns: non-significant (p > 0.05), *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

**Figure 3.**Distribution of group-based, cell-centered MAE scores, calculated across 200 genes that have the highest gene-centered MAE scores. The 200 genes with the highest MAE scores were detected separately for each cell type, and then the cell-centered MAE score was calculated based on these genes. Panel titles indicate the focus groups (for example, the panel entitled ‘H1’ shows the cell-centered MAE score distribution, calculated for each cell type, based on 200 genes that have the highest gene-centered MAE scores within the H1 group). H1: human embryonic stem cells (n = 375 cells); NPC: neural progenitor cells (n = 173 cells); DEP: definitive endoderm progenitors (n = 138 cells); EC: endothelial cells (n = 105 cells); TB: trophoblasts (n = 69 cells); HFF: human foreskin fibroblasts (n = 159 cells). Each group was compared to all other cell types, combined, using a Wilcoxon test. ns: non-significant (p > 0.05), *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

**Figure 4.**Distribution of group-based, cell-centered mean-expression (mean-EXP) scores, calculated across the 200 genes that were found to have the lowest gene-centered mean-EXP scores. The 200 genes with the lowest mean-EXP score values were detected separately for each cell type, and then cell-centered mean-EXP scores were calculated, based on these genes. Panel titles indicate the focus groups. H1: human embryonic stem cells (n = 375 cells); NPC: neural progenitor cells (n = 173 cells); DEP: definitive endoderm progenitors (n = 138 cells); EC: endothelial cells (n = 105 cells); TB: trophoblasts (n = 69 cells); HFF: human foreskin fibroblasts (n = 159 cells). Each group was compared to all other cell types, combined, using a Wilcoxon test. Ns: non-significant (p > 0.05), *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

**Figure 5.**Distribution of group-based, cell-centered polygenic scores (PS), calculated across the 200 most differentially expressed (DE) genes. The 200 DE genes were detected separately for each cell type, and then cell-centered PS scores were calculated, based on these genes. Panel titles indicate the focus group. H1: human embryonic stem cells (n = 375 cells); NPC: neural progenitor cells (n = 173 cells); DEP: definitive endoderm progenitors (n = 138 cells); EC: endothelial cells (n = 105 cells); TB: trophoblasts (n = 69 cells); HFF: human foreskin fibroblasts (n = 159 cells). Each group was compared to all other cell types, combined, using a Wilcoxon test. Ns: non-significant (p > 0.05), *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

**Figure 6.**Visualization of cell-type separation using PCA. Each cell is represented by (

**a**) six MAE values, calculated based on the 200 high-MAE genes of each cell type; (

**b**) six expression values, averaged across the 200 lowest expressed genes for each cell type; or (

**c**) six polygenic scores, calculated based on 200 DE genes of each cell type. H1: human embryonic stem cells (n = 375 cells); NPC: neural progenitor cells (n = 173 cells); DEP: definitive endoderm progenitors (n = 138 cells); EC: endothelial cells (n = 105 cells); TB: trophoblasts (n = 69 cells); HFF: human foreskin fibroblasts (n = 159 cells).

**Figure 7.**Overlap between the 200 high-MAE genes, the 200 low mean-expression (mean-EXP) genes, and 200 differentially expressed (DE) genes of each cell type. (

**a**) Human embryonic stem cells list (H1, n = 375 cells); (

**b**) neural progenitor cells list (NPC, n = 173 cells); (

**c**) definitive endoderm progenitors list (DEC, n = 138 cells); (

**d**) trophoblasts list (TB, n = 69 cells); (

**e**) endothelial cells list (EC, n = 105 cells), and human foreskin fibroblasts list (HFF, n = 159 cells).

**Figure 8.**Distribution of gene-centered MAE scores for human embryonic stem cells (H1). (

**a**) Density plot of MAE scores, calculated separately for each gene (n = 19,097 genes) across all H1 cells (n = 375 cells). (

**b**) Density plot of the expression values of the highest (blue) and lowest (red) MAE genes of the H1 cells.

**Figure 9.**Cell type prediction accuracy of machine-learning algorithms. The area under the ROC curve (AUC), showing the performance of the indicated models in classification of cell types. The models were based on cell scores calculated from the 200 high-MAE (

**a**), low mean-EXP (

**b**), DE (

**c**), Seurat (

**d**), or MAST (

**e**) gene sets.

**Figure 10.**Tissue prediction accuracy of machine-learning algorithms. The area under the ROC curve (AUC), showing the performance of the indicated models in classification of tissue origin. The models were based on cell scores calculated from the 200 high-MAE (

**a**), low mean-EXP (

**b**), or the DE (

**c**) gene sets.

**Table 1.**The dataset from Chu et al. [20], used in this study, indicating the number of single cells within each cell type.

Cell Type | Potency | Number of Cells |
---|---|---|

Human embryonic stem cells (hESC) | Pluripotent | 374 |

Neural progenitor cells (NPC) | Multipotent | 173 |

Definitive endoderm progenitors (DEP) | 138 | |

Endothelial cells (EC) | 105 | |

Trophoblasts (TB) | 69 | |

Human foreskin fibroblasts (HFF) | Differentiated | 159 |

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Morag, S.; Salmon-Divon, M. Characterizing Human Cell Types and Tissue Origin Using the Benford Law. *Cells* **2019**, *8*, 1004.
https://doi.org/10.3390/cells8091004

**AMA Style**

Morag S, Salmon-Divon M. Characterizing Human Cell Types and Tissue Origin Using the Benford Law. *Cells*. 2019; 8(9):1004.
https://doi.org/10.3390/cells8091004

**Chicago/Turabian Style**

Morag, Sne, and Mali Salmon-Divon. 2019. "Characterizing Human Cell Types and Tissue Origin Using the Benford Law" *Cells* 8, no. 9: 1004.
https://doi.org/10.3390/cells8091004