Batch effects and the effective design of single-cell gene expression studies
Open accessPublished version of a preprint
The DOI matches its Crossref record; 57 of the 57 references it cites were found in a registry or an index; 4 of the 18 sources its claims cite were checked for support.
Abstract
Single-cell RNA sequencing (scRNA-seq) can be used to characterize variation in gene expression levels at high resolution. However, the sources of experimental noise in scRNA-seq are not yet well understood. We investigated the technical variation associated with sample processing using the single-cell Fluidigm C1 platform. To do so, we processed three C1 replicates from three human induced pluripotent stem cell (iPSC) lines. We added unique molecular identifiers (UMIs) to all samples, to account for amplification bias. We found that the major source of variation in the gene expression data was driven by genotype, but we also observed substantial variation between the technical replicates. We observed that the conversion of reads to molecules using the UMIs was impacted by both biological and technical variation, indicating that UMI counts are not an unbiased estimator of gene expression levels. Based on our results, we suggest a framework for effective scRNA-seq studies.
Overview
CoChat Summary
TL;DR
Using three human iPSC lines processed across multiple Fluidigm C1 replicates with ERCC spike-ins and UMIs, this study finds that genotype is the largest source of expression variation but that C1 batch effects are substantial and affect both endogenous genes and ERCC controls. The authors recommend balanced mixed-individual C1 designs and at least 75 cells per condition at ≥1.5 million reads per cell, and show that ERCC spike-ins cannot fully correct batch effects.
Key findings
- Individual/genotype was the major source of variation, but C1 batch effects were also substantial: median variance due to individual was 8% versus 4% due to C1 batch (Kruskal-Wallis, P < 0.001), and cells within a batch were more correlated than cells from the same individual in different batches (P < 0.001).
- ERCC spike-ins were affected by both technical and biological variation and could not effectively account for C1 batch effects: total ERCC molecule counts differed between C1 batches (F-test, P < 0.001) and across individuals (LRT, P = 0.08); per-gene ERCC variance was larger for individual (median 9%) than batch (median 5%; P < 0.001). Read-to-molecule conversion efficiency showed batch effects for ERCC (P < 0.05) and endogenous genes (P < 0.001), and individual effects for endogenous (P < 0.01) and ERCC (P < 0.01).
- UMI-based molecule counts were not an unbiased estimator of gene expression: read and molecule counts correlated at r = 0.99 for ERCC spike-ins but r = 0.92 for endogenous genes, with lower correlation for lowly expressed genes.
- After batch correction, 560 genes had significantly different regulatory noise across individuals (empirical P < 0.0001), enriched for translation, protein disassembly, and biosynthetic processes; among them were pluripotency genes KLF4 and DPPA2. Only 95 of the 1,000 most highly variable genes were common to all three individuals, whereas 887 of the 1,000 most highly expressed genes were common.
Methods
- scRNA-seq was performed on three Yoruba iPSC lines using the Fluidigm C1 platform, with three independent C1 collections per individual, matching bulk samples, ERCC spike-in controls, and 5-bp random-sequence UMIs. After quality control, 564 high-quality single cells were retained across eight replicates from three individuals (NA19098: 142, NA19101: 201, NA19239: 221).
- Reads were mapped, filtered, and converted to molecule counts using UMI-tools; analyses included PCA, nested linear mixed models with individual fixed effects and batch random effects, variance component estimation, likelihood ratio tests, F-tests, Kruskal-Wallis tests, and subsampling of cells and reads.
- Normalization included UMI collision correction, log2 counts-per-million transformation, ERCC-based Poisson GLM correction, and batch-effect removal using a limma-like linear mixed model. Regulatory noise was measured as adjusted coefficient of variation, with differential noise genes identified by median absolute deviation and permutation (300,000 permutations).
Limitations
- Stated: the batch correction assumes the technical batch effect is identical across individuals with no interaction, and the small number of individuals prevented explicit testing of a first-order batch-by-individual interaction.
- Stated: ERCC controls do not experience all processing steps of endogenous samples, have shorter polyA tails and are shorter than mammalian mRNAs, only about half of the 92 controls are typically detected, and they are affected by biological as well as technical variation; the authors recommend an alternative control such as Drosophila RNA for human single-cell studies.
- Evident from the text: the study is a small pilot with only three individuals, one C1 replicate was excluded because of a pipetting error, and the conclusions about study design and noise are based on this limited sample.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures





Verification
Claims the paper makes
3 Valid · 0 Not valid · 7 Not checked · 88 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“One of the biggest limitations of using the C1 system, however, is that single cell capture and preparation from different conditions are fully independent23.”
§ Batch effects and the effective design of single-cell gene expression studies > Abstract
What the sources say
- Partly supportedcites [23] Hicks, S. C., Teng, M. & Irizarry, R. A. et al. On the widespread and critical impact of systematic bias and batch effects in single-cell RNA-seq data. bioRxiv doi: 10.1101/025528 (2015). sourceExperimental — accuracy not yet measured
Our copy states the underlying point in general terms: batch effects arise 'when cells from one biological group or condition are cultured, captured and sequenced separate from cells in a second condition' and 'due to the nature of certain experimental scRNA-seq protocols, which restrict the way cells are captured and sequenced separately, sometimes standard balanced experimental designs are not possible' (Introduction), and the Discussion calls this 'experimental limitation' a challenge that makes it 'difficult to impossible to parse technical from biological variation' (p. 4). However, the paper never names the C1/Fluidigm platform (searched: Fluidigm, C1, microfluidic; Table 1 lists single-cell technologies only in a truncated table caption), so the attribution of this as 'one of the biggest limitations of the C1 system' specifically is not supported by our copy — the source speaks about scRNA-seq protocols in general, not about C1. (LLM, conf=0.72)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“We computed a measure of gene expression noise that is independent of RNA abundance across individuals4647.”
§ Batch effects and the effective design of single-cell gene expression studies > Methods > Measurement of gene expression noise
What the sources say
- Supportedcites [46] Kolodziejczyk, K., A. A. & Marioni, J. C. et al. Single cell RNA-sequencing of pluripotent states unlocks modular transcriptional variation. Cell Stem Cell 17, 471–485 (2015). source
In Results, under “Transcriptome-wide Cell-to-Cell Variation Is Similar across the Three Culture Conditions,” the paper says its distance-to-the-median measure accounts for the dependence of gene-expression variability on mean expression level. This supports the claim that the computed noise measure is normalized for RNA abundance across the individual cells studied. (LLM, conf=0.88)
- Not checked yetcites [47] Newman, G., J. R. S. & Weissman, J. S. et al. Single-cell proteomic analysis of s. cerevisiae reveals the architecture of biological noise. Nature 441, 840–846 (2006). source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“at this level of subsampling for individual NA19239, we were able to detect a mean of 6068 genes out of 6097 genes expressed in the bulk samples (the bottom 50%; Fig. 2b)”
§ Batch effects and the effective design of single-cell gene expression studies > Results > Study design and quality control
plainSingle case- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Indeed, we observed that total ERCC molecule-counts are significantly different between C1 batches (F-test; P < 0.001).”
§ Batch effects and the effective design of single-cell gene expression studies > Results > Batch effects associated with UMI-based single cell data
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“but we also observed substantial variation between the technical replicates.”
- Also stated in: § Batch effects associated with UMI-based single cell data · plain
“Yet, variation due to C1 batch is also substantial - data from single cell samples within a batch are more correlated than that from single cells from the same individual but different batches”
§ Batch effects and the effective design of single-cell gene expression studies > Abstract
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Also stated in: § Batch effects associated with UMI-based single cell data · plain
70 more claims
Create a free account to see all 98 claims and to verify the 14 references still open.
- It may state two separate claims as one. 10
- Probably not a substantial claim. 14
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The presence of variability between replicates is unavoidable, and can be attributed to both technical factors and intrinsic biological differences 22 .”
From Confronting false discoveries in single-cell differential expression (2021) · cites this paper as [22]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Increasing biological replicates through sample pooling prior to library preparation followed by demultiplexing (i.e., identifying and isolating the sequencing data related to each sample within the pool) to the original biological sample is an effective strategy to overcome these issues as well as reduce batch effects [5–9].”
From Benchmarking of computational demultiplexing methods for single-nucleus RNA sequencing data (2025) · cites this paper as [6]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Within the same experimental condition, replicates exhibit inherent differences in gene expression, which reflect both biological and technical factors 22 .”
From Confronting false discoveries in single-cell differential expression (2021) · cites this paper as [22]
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“The analysis of multiplexed dscRNA-seq data recovers cell type-specific gene regulatory programs affected by interferon stimulation consistent with published IFN-β signatures in mouse and humans29.”
From Multiplexed droplet single-cell RNA-sequencing using natural genetic variation (2018) · cites this paper as [29]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
“We found significant individual differences (LRT; P < 10−5) in gene-specific dropout rates between individuals in more than 10% (1,214 of 13,058) of expressed endogenous genes.”
§ Batch effects and the effective design of single-cell gene expression studies > Results > Measuring regulatory noise in single-cell gene expression data
Reported 10 · recomputed 9.296983
Citations
Cites
The Sequence Alignment/Map format and SAMtools
Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth, G., Abecasis, G., Durbin, R., & 1000 Genome Project Data Processing Subgroup (2009) · Bioinformatics · doi:10.1093/bioinformatics/btp352 · Cited by 62,041 · cited 1×
“First, we combined all reads for a given 404 single cell using samtools [42].”
Methodlimma powers differential expression analyses for RNA-sequencing and microarray studies
Ritchie, M. E., Phipson, B., Wu, D., Hu, Y., Law, C. W., Shi, W., & Smyth, G. K. (2015) · Nucleic Acids Research · doi:10.1093/nar/gkv007 · Cited by 40,458 · cited 1×
“series function of the limma package [45].”
featureCounts: an efficient general purpose program for assigning sequence reads to genomic features
Liao, Y., Smyth, G. K., & Shi, W. (2014) · Bioinformatics · doi:10.1093/bioinformatics/btt656 · Cited by 27,219 · cited 1×
“To obtain gene-level counts, we assigned reads to protein-coding genes (Ensembl 397 GRCh37 release 82) and the ERCC spike-in genes using featureCounts [41].”
MethodGene Expression Omnibus: NCBI gene expression and hybridization array data repository
Edgar, R. (2002) · Nucleic Acids Research · doi:10.1093/nar/30.1.207 · Cited by 13,288
An integrated map of genetic variation from 1,092 human genomes
The 1000 Genomes Project Consortium (2012) · Nature · doi:10.1038/nature11632 · Cited by 7,711
Highly parallel genome-wide expression profiling of individual cells using nanoliter droplets.
Macosko, E. Z., Basu, A., Satija, R., Nemesh, J., Shekhar, K., Goldman, M., Tirosh, I., Bialas, A. R., Kamitaki, N., Martersteck, E. M., et al. (2015) · Cell · doi:10.1016/j.cell.2015.05.002 · Cited by 7,110 · cited 2×
“The automated C1 microfluidic 52 platform (Fluidigm), while more expensive per sample, has been shown to confer several 53 advantages over platforms that make use of droplets to capture single cells [3,21].”
BackgroundResultA haplotype map of the human genome
The International HapMap Consortium (2005) · Nature · doi:10.1038/nature04226 · Cited by 5,858
Droplet barcoding for single-cell transcriptomics applied to embryonic stem cells
Klein, A. M., Mazutis, L., Akartuna, I., Tallapragada, N., Veres, A., Li, V., Peshkin, L., Weitz, D. A., & Kirschner, M. W. (2015) · Cell · doi:10.1016/j.cell.2015.04.044 · Cited by 3,356
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. (2015) · Genome Biology · doi:10.1186/s13059-015-0844-5 · Cited by 3,159
The Subread aligner: fast, accurate and scalable read mapping by seed-and-vote
Liao, Y., Smyth, G. K., & Shi, W. (2013) · Nucleic Acids Research · doi:10.1093/nar/gkt214 · Cited by 3,064 · cited 2×
“We then mapped reads to human genome hg19 (only including chromosomes 1-22, X, and Y, plus the ERCC sequences) with Subjunc [40], discarding non-uniquely mapped reads (option-u).”
Cited by
Batch effects in single-cell RNA-sequencing data are corrected by matching mutual nearest neighbors
Haghverdi, L., Lun, A. T. L., Morgan, M. D., & Marioni, J. C. (2018) · Nature Biotechnology · doi:10.1038/nbt.4091 · Cited by 2,416
“All of these factors result in batch effects [5] [6], where the expression of genes in one batch differs systematically from those in another batch.”
Cao, J., Packer, J. S., Ramani, V., Cusanovich, D. A., Huynh, C., Daza, R., Qiu, X., Lee, C., Furlan, S. N., Steemers, F. J., et al. (2017) · doi:10.1101/104844 · Cited by 1,402
Accelerating Medicines Partnership Rheumatoid Arthritis and Systemic Lupus Erythematosus (AMP RA/SLE) Consortium, Zhang, F., Wei, K., Slowikowski, K., Fonseka, C. Y., Rao, D. A., Kelly, S., Goodman, S. M., Tabechian, D., Hughes, L. B., et al. (2019) · Nature Immunology · doi:10.1038/s41590-019-0378-1 · Cited by 1,367
A single-cell transcriptomic atlas characterizes ageing tissues in the mouse
The Tabula Muris Consortium, Almanzar, N., Antony, J., Baghel, A. S., Bakerman, I., Bansal, I., Barres, B. A., Beachy, P. A., Berdnik, D., Bilen, B., et al. (2020) · Nature · doi:10.1038/s41586-020-2496-1 · Cited by 1,247
Multiplexed droplet single-cell RNA-sequencing using natural genetic variation
Kang, H. M., Subramaniam, M., Targ, S., Nguyen, M., Maliskova, L., McCarthy, E., Wan, E., Wong, S., Byrnes, L., Lanata, C. M., et al. (2018) · Nature Biotechnology · doi:10.1038/nbt.4042 · Cited by 1,101
Metadata
- Authors
- Tung, P.-Y., Blischak, J. D., Hsiao, C. J., Knowles, D. A., Burnett, J. E., Pritchard, J. K., & Gilad, Y.
- Affiliations
- University of Chicago (US)
- Stanford Medicine (US)
- Howard Hughes Medical Institute (US)
- Corresponding author
- Gilad, Yoav
- Venue
- Scientific Reports
- Published
- 2017-01-03
- Volume
- 7
- Issue
- 1
- Pages
- 39921-39921
- ISSN
- 2045-2322 (ISSN-L)
- PMC
- PMC5206706
- OpenAlex
- W2951638683
- Cited by
- 356
- References
- 55
- Citations by year
- 2026: 23 · 2025: 35 · 2024: 32 · 2023: 22 · 2022: 38 · 2021: 41 · 2020: 52 · 2019: 42 · 2018: 36 · 2017: 31
- FWCI
- 18.09
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by
- creativecommons.org/licenses/by/4.0 (tdm)
- creativecommons.org/licenses/by/4.0 (vor)
- Locations
- Scientific Reports (publishedVersion, cc-by)
- bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Funders
- Howard Hughes Medical Institute
- National Institutes of Health
- National Heart, Lung, and Blood Institute (R01HL092206, T32HL007381)
- National Institute of General Medical Sciences (T32GM007197)
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Single-cell and spatial transcriptomics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene expression and cancer classification
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
- Keywords
- single-cell RNA sequencing, batch effects, technical variability, unique molecular identifiers, amplification bias, gene expression quantification, induced pluripotent stem cells, genotypic variation, research design
- MeSH terms
- Humans, RNA, Gene Expression, Sequence Analysis, RNA, Principal Component Analysis, Induced Pluripotent Stem Cells, Single-Cell Analysis, High-Throughput Nucleotide Sequencing
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-04, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-04 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (qwen3.8-omni-flash and gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Batch effects and the effective design of single-cell gene expression studies. https://app.cochat.ai/papers/10.1038/srep39921
- The paper
- Tung, P.-Y., Blischak, J. D., Hsiao, C. J., Knowles, D. A., Burnett, J. E., Pritchard, J. K., & Gilad, Y. (2017). Batch effects and the effective design of single-cell gene expression studies. Scientific Reports. https://doi.org/10.1038/srep39921
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.