Scaling Laws for Masked-Reconstruction Transformers on Single-Cell Transcriptomics

• URL: http://arxiv.org/abs/2602.15253v1
• Date: Mon, 16 Feb 2026 23:20:58 GMT
• Title: Scaling Laws for Masked-Reconstruction Transformers on Single-Cell Transcriptomics
• Authors: Ihor Kendiukhov,
• Abstract summary: We present the first systematic study of scaling behaviour for masked-reconstruction transformers trained on single-cell RNA sequencing data.<n>Across seven model sizes spanning three orders of magnitude in parameter count, we fit the parametric scaling law to validation mean squared error (MSE)<n>Results establish that scaling laws analogous to those observed in natural language processing do emerge in single-cell transcriptomics when sufficient data are available.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Neural scaling laws -- power-law relationships between loss, model size, and data -- have been extensively documented for language and vision transformers, yet their existence in single-cell genomics remains largely unexplored. We present the first systematic study of scaling behaviour for masked-reconstruction transformers trained on single-cell RNA sequencing (scRNA-seq) data. Using expression profiles from the CELLxGENE Census, we construct two experimental regimes: a data-rich regime (512 highly variable genes, 200,000 cells) and a data-limited regime (1,024 genes, 10,000 cells). Across seven model sizes spanning three orders of magnitude in parameter count (533 to 3.4 x 10^8 parameters), we fit the parametric scaling law to validation mean squared error (MSE). The data-rich regime exhibits clear power-law scaling with an irreducible loss floor of c ~ 1.44, while the data-limited regime shows negligible scaling, indicating that model capacity is not the binding constraint when data are scarce. These results establish that scaling laws analogous to those observed in natural language processing do emerge in single-cell transcriptomics when sufficient data are available, and they identify the data-to-parameter ratio as a critical determinant of scaling behaviour. A preliminary conversion of the data-rich asymptotic floor to information-theoretic units yields an estimate of approximately 2.30 bits of entropy per masked gene position. We discuss implications for the design of single-cell foundation models and outline the additional measurements needed to refine this entropy estimate.

Related papers

• Unlasting: Unpaired Single-Cell Multi-Perturbation Estimation by Dual Conditional Diffusion Implicit Bridges [68.98973318553983]
  We propose a framework based on Dual Diffusion Implicit Bridges (DDIB) to learn the mapping between different data distributions.<n>We integrate gene regulatory network (GRN) information to propagate perturbation signals in a biologically meaningful way.<n>We also incorporate a masking mechanism to predict silent genes, improving the quality of generated profiles.
  arXiv  Detail & Related papers  (2025-06-26T09:05:38Z)
• A scalable gene network model of regulatory dynamics in single cells [88.48246132084441]
  We introduce a Functional Learnable model of Cell dynamicS, FLeCS, that incorporates gene network structure into coupled differential equations to model gene regulatory functions.<n>Given (pseudo)time-series single-cell data, FLeCS accurately infers cell dynamics at scale.
  arXiv  Detail & Related papers  (2025-03-25T19:19:21Z)
• Training Compute-Optimal Protein Language Models [48.79416103951816]
  Most protein language models are trained with extensive compute resources until performance gains plateau. Our investigation is grounded in a massive dataset consisting of 939 million protein sequences. We trained over 300 models ranging from 3.5 million to 10.7 billion parameters on 5 to 200 billion unique tokens.
  arXiv  Detail & Related papers  (2024-11-04T14:58:37Z)
• Large-Scale Targeted Cause Discovery via Learning from Simulated Data [66.51307552703685]
  We propose a novel machine learning approach for inferring causal variables of a target variable from observations.<n>We train a neural network using supervised learning on simulated data to infer causality.<n> Empirical results demonstrate superior performance in identifying causal relationships within large-scale gene regulatory networks.
  arXiv  Detail & Related papers  (2024-08-29T02:21:11Z)
• sc-OTGM: Single-Cell Perturbation Modeling by Solving Optimal Mass Transport on the Manifold of Gaussian Mixtures [0.9674145073701153]
  sc-OTGM is an unsupervised model grounded in the inductive bias that the scRNAseq data can be generated. sc-OTGM is effective in cell state classification, aids in the analysis of differential gene expression, and ranks genes for target identification. It also predicts the effects of single-gene perturbations on downstream gene regulation and generates synthetic scRNA-seq data conditioned on specific cell states.
  arXiv  Detail & Related papers  (2024-05-06T06:46:11Z)
• xTrimoGene: An Efficient and Scalable Representation Learner for Single-Cell RNA-Seq Data [45.043516102428676]
  We propose a novel asymmetric encoder-decoder transformer for scRNA-seq data, called xTrimoGene$alpha$ (or xTrimoGene for short) xTrimoGene reduces FLOPs by one to two orders of magnitude compared to classical transformers while maintaining high accuracy. Our experiments also show that the performance of xTrimoGene improves as we scale up the model sizes.
  arXiv  Detail & Related papers  (2023-11-26T01:23:01Z)
• A Solvable Model of Neural Scaling Laws [72.8349503901712]
  Large language models with a huge number of parameters, when trained on near internet-sized number of tokens, have been empirically shown to obey neural scaling laws. We propose a statistical model -- a joint generative data model and random feature model -- that captures this neural scaling phenomenology. Key findings are the manner in which the power laws that occur in the statistics of natural datasets are extended by nonlinear random feature maps.
  arXiv  Detail & Related papers  (2022-10-30T15:13:18Z)
• Modelling Technical and Biological Effects in scRNA-seq data with Scalable GPLVMs [6.708052194104378]
  We extend a popular approach for probabilistic non-linear dimensionality reduction, the Gaussian process latent variable model, to scale to massive single-cell datasets. The key idea is to use an augmented kernel which preserves the factorisability of the lower bound allowing for fast variational inference.
  arXiv  Detail & Related papers  (2022-09-14T15:25:15Z)
• Explaining Neural Scaling Laws [17.115592382420626]
  Population loss of trained deep neural networks often follows precise power-law scaling relations. We propose a theory that explains the origins of and connects these scaling laws. We identify variance-limited and resolution-limited scaling behavior for both dataset and model size.
  arXiv  Detail & Related papers  (2021-02-12T18:57:46Z)
• A Systematic Approach to Featurization for Cancer Drug Sensitivity Predictions with Deep Learning [49.86828302591469]
  We train >35,000 neural network models, sweeping over common featurization techniques. We found the RNA-seq to be highly redundant and informative even with subsets larger than 128 features.
  arXiv  Detail & Related papers  (2020-04-30T20:42:17Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.