A hybrid demultiplexing strategy that improves performance and robustness of cell hashing
Bibliographic record
Abstract
Abstract Recent advances in single cell RNA sequencing allow users to pool multiple samples and demultiplex in downstream analysis, which greatly increase experimental efficiency and cost-effectiveness. Among all the demultiplexing methods, nucleotide barcode-based cell hashing has gained widespread popularity due to its compatibility and simplicity. Despite these advantages, certain issues of this technic remain to be solved, such as challenges in distinguishing true positive from background, high reagent cost for samples with large cell numbers, and unpredictable false negative and false doublet rates. Here, we propose a hybrid demultiplexing strategy that increases calling accuracy and cell recovery of cell hashing without adding experimental cost. In this approach, we computationally cluster all single cells based on their natural genetic variations and assign donor identity by finding the dominant hashtag in each genotype cluster. This hybrid strategy assigns donor identity to any cell that is identified as singlet by either genotype clustering or cell hashing, which allows us to demultiplex most majority of cells even if only a small fraction of cells are labeled with hashtags. When comparing its performance with cell hashing on multiple real-world datasets, this hybrid approach consistently generates reliable demultiplexing results with increased cell recovery and accuracy. Key Points The improved cut-off calling tool, HTOreader, accurately distinguishes true positive from background signal for each individual hashtag. The hybrid demultiplexing strategy increases cell recovery of cell hashing by increasing cut-off calling accuracy and decreasing false negative and false double rates. The hybrid strategy enhances cost-effectiveness of cell hashing and consistently produces reliable demultiplexing results, regardless of hashtag staining quality. The hybrid strategy can be seamlessly integrated into a variety of single-cell experimental protocols and analytic pipelines.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.001 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
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".