DECONVersation 0.0.1

DECONVersation is a tool designed for the deconvolution of bulk RNA-seq data using embeddings derived from large-scale, LLM-based foundation models. DECONVersation produces robust cell type proportions by leveraging these high-dimensional embeddings to mitigate batch effects typically present in single-cell reference signature matrices.

DECONVersation leverages embedding representations from large-scale, LLM-based foundation models to perform deconvolution of bulk RNA-seq data. Currently, embeddings from Geneformer, Cell2Sentence, CellHermes, and scGPT are supported.

---
config:
  theme: 'neutral'
---
flowchart
    subgraph ide1 [standard]
    direction TB 
    A[scRNA: cell x gene] --> B(full profile: type x gene)
    A[scRNA: cell x gene] --> C(markers)
    D[bulkRNA: sample x gene]
    B --> F(signature: type x marker)
    C --> F
    F bb@==> E
    D db@==> E{{deconv res:
    sample x type%}}
    end
    subgraph ide2 [foundation model]
    direction TB
    A1[scRNA: cell x gene] --> B1(full profile: type x gene)
    A1[scRNA: cell x gene] --> C1([fa:fa-robot finetuned model])
    B1 --> F1(type x embeddings)
    C1 --> F1
    D1[bulkRNA: sample x gene] --> G1(sample x embeddings)
    C1 --> G1
    F1 f1b@==> E1{{deconv res:
    sample x type%}}
    G1 g1b@==> E1
    end

bb@{ curve: linear }
db@{ curve: linear }
f1b@{ curve: linear }
g1b@{ curve: linear }
style A fill:green,color:#fff
style A1 fill:green,color:#fff
style D fill:blue,color:#fff
style D1 fill:blue,color:#fff
style C1 fill:red,color:#fff
style E stroke-width:4px
style E1 stroke-width:4px

---

Overview

This project provides:

• Installation guide for DECONVersation
• Step-by-step tutorials for embedding extraction and downstream deconvolution analysis
• Sample pseudobulk dataset for deconvolution testing
• Perfomance evaluation comparing estimates from DECONVersation and other deconvolution tools
• Guide to finetuning foundational models using DECONVersation (Geneformer and Cell2Sentence)

---

DECONVersation Features

DECONVersation supports end-to-end deconvolution through a set of easy-to-use functions. Geneformer, Cell2Sentence, CellHermes, and scGPT embeddings can be extracted from both bulk and single-cell datasets, with single-cell embeddings used to construct robust signature matrices from .h5ad references. Cell type proportions are then estimated via NNLS directly in embedding space. Built-in benchmarking tools evaluate predictions against ground truth using RMSE and Pearson correlation, complemented by visualization utilities for assessing method performance. DECONVersation also supports testing and validation with in-built pseudobulk functions.

---

Tutorials

• DECONVersation on bulk RNA-seq using Geneformer: How to extract embeddings (using geneformer)and run DECONVersation on bulk using a single cell reference.
• DECONVersation on pseudobulk using Geneformer: Validate deconvolution using pseudobulk data

---

Benchmarking

DECONVersation was benchmarked across 6 real bulk RNA-seq datasets with ground truths, spanning diverse tissue types and experimental conditions, to evaluate deconvolution performance and generalizability.

Summary
Across 6 benchmarked datasets, we show overall RMSE and correlation coefficient alongside mean RMSE and correlation averaged across cell types. Finetuned Cell2Sentence and Geneformer-based embeddings both demonstrate consistent deconvolution performance across all 6 datasets, with finetuned models outperforming their zero-shot counterparts in each case. This highlights the importance of finetuning. Finetuning was achieved by training models to predict cell type annotations from a single-cell reference. Among the compared methods, only DWLS achieves comparable performance to the finetuned embedding-based approaches available on DECONVersation.

#	Dataset	Source	Ground Truth	Cell Type #
1	PMBC (Hoek)	PBMC	FACS	5
2	PMBC (Finotello)	PBMC	FACS	5
3	PMBC (Morandini)	PBMC	FACS	5
4	Cell Line Mixture (Cobos)	Cell Line Mixture	Mixture Count	6
5	Pre-Frontal Cortex (Huuki-Myers)	DLPFC	RNAScope/IF	6
6	Retina (Guo)	Retina	snRNA	6

---

Suggested Reading

• Geneformer Transfer learning enables predictions in network biology
• Cell2Sentence Cell2Sentence: Teaching Large Language Models the Language of Biology
• CellHermes Language may be all omics needs: Harmonizing multimodal data for omics understanding with CellHermes
• scGPT scGPT: toward building a foundation model for single-cell multi-omics using generative AI

---