MétaCan
Menu
Back to cohort
Record W4414874842 · doi:10.1101/2025.10.06.680609

A Fully Automated, Data-Driven Approach for Dimensionality Reduction and Clustering in Single-Cell RNA-seq Analysis

2025· preprint· en· W4414874842 on OpenAlexaff
Hyung Ham Kim, Faeyza Rishad Ardi, Kévin Spinicci, Jae Kyoung Kim

Bibliographic record

VenuebioRxiv (Cold Spring Harbor Laboratory) · 2025
Typepreprint
Languageen
FieldBiochemistry, Genetics and Molecular Biology
TopicSingle-cell and spatial transcriptomics
Canadian institutionsKootenay Association for Science & Technology
FundersInstitute for Basic Science
KeywordsCluster analysisPrincipal component analysisPipeline (software)Dimensionality reductionNoise (video)Reliability (semiconductor)Curse of dimensionalityRand indexClustering high-dimensional data

Abstract

fetched live from OpenAlex

Abstract Single-cell RNA sequencing (scRNA-seq) provides deep insights into cellular heterogeneity but demands robust dimensionality reduction (DR) and clustering to handle high-dimensional, noisy data. Many DR and clustering approaches rely on user-defined parameters, undermining reliability. Even automated clustering methods like ChooseR and MultiK still employ fixed principal component defaults, limiting their full automation. To overcome this limitation, we propose a fully automated clustering approach by integrating scLENS—a method for optimal PC selection—with these tools. Our fully automated approach improves clustering performance by ∼14% for ChooseR and ∼10% for MultiK and identifies additional cell subtypes, highlighting the advantages of adaptive, data-driven DR. Highlights Fully automated, data-driven clustering pipeline for scRNA-seq analysis. Addresses the limitation of fixed principal component defaults in DR. Data-driven pipeline improves clustering performance by 10-14%. Performance gains are most pronounced on high-sparsity, high-skewness data. Author Summary A fundamental goal in modern biology is to create detailed cellular maps of complex tissues to understand their function in health and disease. Achieving this level of detail is now possible through single-cell sequencing, a technology that generates massive, high-dimensional, and noisy datasets of individual cells’ genetic profiles. However, accurately identifying cell-types from these datasets remains a major analytical hurdle, as current computational methods rely on subjective parameters that compromise reliability and reproducibility. To address this challenge, we developed a fully automated pipeline that integrates scLENS, an adaptive noise filtering method, with automated cell grouping tools. By providing optimal parameters for noise filtering and cell grouping, our pipeline eliminates their reliance on subjective, fixed parameters, thereby significantly improving accuracy in cell-type classification. This tool, characterized by its high clustering efficiency, can accelerate discoveries in fields like cancer biology and immunology by providing a robust and reproducible analytical platform.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame machine prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.

metaresearch head score (Codex)0.003
metaresearch head score (Gemma)0.006
Version: metacan-v3-hybrid-931329e0061cValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: none
GenreCandidate signal: Methods · Consensus signal: Methods
Teacher disagreement score0.005
Threshold uncertainty score0.016

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.0030.006
Meta-epidemiology (narrow)0.0020.001
Meta-epidemiology (broad)0.0020.003
Bibliometrics0.0020.002
Science and technology studies0.0020.001
Scholarly communication0.0030.001
Open science0.0030.003
Research integrity0.0010.003
Insufficient payload (model declined to judge)0.0050.006

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.028
GPT teacher head0.245
Teacher spread0.217 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.

The models applied no category: nothing in the taxonomy fit this work.
Study designSimulation or modeling
Domainnot available
GenreMethods

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations0
Published2025
Admission routes1
Has abstractyes

Explore more

Same venuebioRxiv (Cold Spring Harbor Laboratory)Same topicSingle-cell and spatial transcriptomicsFrench-language works237,207