Biology subjects

Litman, E.

Publications and source records attributed to Litman, E..

4 recordsLinked to original sources

bioRxiv2026

MetaMuse: A Multi-Agent AI System for Biomedical Metadata Curation and Harmonization

Inconsistent and unstructured metadata in public biomedical repositories, such as the Gene Expression Omnibus (GEO), severely limits data discoverability and research reproducibility. To address this, we introduce MO_SCPLOWETAC_SCPLOWMO_SCPLOWUSEC_SCPLOW, a modular, multi-agent artificial intelligence framework designed to autonomously extract, validate, and standardize unstructured biomedical metadata. Operating through a three-stage architecture utilizing large language model agents, specialized CO_SCPLOWURATORC_SCPLOWAO_SCPLOWGENTSC_SCPLOW contextually extract candidate values for specific target metadata fields. A centralized AO_SCPLOWRBITRATORC_SCPLOWAO_SCPLOWGENTC_SCPLOW enforces cross-field logical consistency to prevent contradictory annotations. Finally, a NO_SCPLOWORMALIZERC_SCPLOWAO_SCPLOWGENTC_SCPLOW leveraging a domain-specific semantic search model (SapBERT) maps these free-text candidates to formal ontological terms. We evaluated MO_SCPLOWETAC_SCPLOWMO_SCPLOWUSEC_SCPLOW on a gold-standard dataset of manually curated GEO samples, achieving over 95% curation accuracy across key target metadata fields, and demonstrated robust scalability on a broader dataset of 400 samples. Notably, MO_SCPLOWETAC_SCPLOWMO_SCPLOWUSEC_SCPLOW avoids data hallucination by defaulting to conservative false negatives when evidence is ambiguous, thereby preserving strict data integrity. By providing a fully auditable and context-aware curation pipeline, MO_SCPLOWETAC_SCPLOWMO_SCPLOWUSEC_SCPLOW offers a scalable solution for enriching public data repositories and accelerating reproducible, data-driven scientific discovery.

genomics↗

bioRxiv2025

GeneJEPA: A Predictive World Model of the Transcriptome

We introduce GO_SCPLOWENEC_SCPLOWJO_SCPLOWEPAC_SCPLOW, a self-supervised foundation model that learns a predictive world model of single-cell transcriptomes. Based on the Joint-Embedding Predictive Architecture, GO_SCPLOWENEC_SCPLOWJO_SCPLOWEPAC_SCPLOW predicts latent representations of masked gene sets from visible context, a shift away from reconstructing noisy expression values and toward world-model style inference over cellular state. To realize this at scale, a Perceiver encoder handles variable gene sets at fixed cost, and a tokenizer jointly represents gene identity and continuous expression using Fourier features. Trained on the Tahoe-100M atlas, GO_SCPLOWENEC_SCPLOWJO_SCPLOWEPAC_SCPLOW learns general representations that transfer across tissues and datasets. On downstream tasks, including drug response and perturbation prediction, it surpasses strong baselines and enables test-time scaling by progressively enlarging the cross-attention over the gene set, trading a small read cost for higher accuracy at inference. GO_SCPLOWENEC_SCPLOWJO_SCPLOWEPAC_SCPLOW moves toward foundation models that reason over gene-gene relations, enabling applications in annotation, prediction, and in-silico discovery.

genomics↗

bioRxiv2025

REVIVE-Flow: A Foundation Model for Blood DNAm Aging

Epigenetic clocks can predict biological age but cannot prescribe the interventions needed to reverse it. Here, we introduce REjuVenatIon Via Epigenetic Flow (Revive-Flow), a flow-matching model trained on a broad compendium of epigenetic blood studies to transport methylomes forward and backward in time. First, we learn the continuous vector field of aging as an Ordinary Differential Equation (ODE) within a stable, low-dimensional linear space. Then, the learned ODEs dynamics are integrated backward in time to define a natural, biologically-plausible rejuvenation trajectory. This path serves as a guide for a convex optimization problem that identifies the minimal, targeted CpG-level perturbation required to rejuvenate a sample. On a completely unseen test set comprising over 800 individuals from the European Prospective Investigation into Cancer and Nutrition (EPIC-Italy) cohort, Revive achieves 0.4 years rejuvenated per commanded year (R2 {approx} 0.99), with a smooth sparsity-effect trade-off. Extensive validation confirms the effect preserves inferred cell-type composition and targets biologically plausible loci enriched in genomic regions and pathways central to aging biology.

genomics↗

bioRxiv2025

Melchior: A Hybrid Mamba-Transformer RNA Basecaller

AO_SCPLOWBSTRACTC_SCPLOWSequence transduction from raw nanopore signals is notoriously difficult because the signal level does not naturally correspond to a single base, but rather many adjacent nucleotides. Thus, we introduce Melchior, an RNA basecaller that uses a hybrid Mamba-Transformer backbone to achieve global visual context at a lower computational complexity. This is in contrast to temporal convolutions, which collate features by fusing both spatial and channel features in the local receptive field. Melchior is also able to exploit the full complementarity between local and global features, unlike Vision Transformers, which have empirically been observed to ignore local features. Augmenting a selective structured state-space sequence model with self-attention unlocks unprecedented performance gains, particularly in homopolymer regions, by modeling fine-grained details in both short and long-range spatial dependencies.

bioinformatics↗