---
title: "ComBat outperforms alternatives for cross-batch prediction in B‑cell lymphoma harmonization"
id: "biorxiv-14-2-234-attempts-to-harmonize-a-multi-platform-b-cell-lymphoma-bulk"
canonical_url: "https://medichelpline.com/clinical-feed/biorxiv-14-2-234-attempts-to-harmonize-a-multi-platform-b-cell-lymphoma-bulk"
content_type: "clinical_feed_article"
specialty: "Oncology"
source_name: "bioRxiv (Biomedical Preprints)"
source_url: "https://www.biorxiv.org/content/10.64898/2026.10.08.757652v1?rss=1"
published_at: "2026-10-15T12:00:00.000Z"
evidence_level: "Verified Feed"
license: "CC-BY-NC-4.0 / Informational Use"
---
# ComBat outperforms alternatives for cross-batch prediction in B‑cell lymphoma harmonization
## Provenance & Clinical Metadata
- **Canonical URL:** https://medichelpline.com/clinical-feed/biorxiv-14-2-234-attempts-to-harmonize-a-multi-platform-b-cell-lymphoma-bulk
- **Specialty:** [Oncology](https://medichelpline.com/clinical-feed/oncology.md)
- **Primary Source:** bioRxiv (Biomedical Preprints)
- **Source URL:** [Original Journal Publication](https://www.biorxiv.org/content/10.64898/2026.10.08.757652v1?rss=1)
- **Published At:** 2026-10-15T12:00:00.000Z
- **Evidence Rating:** Verified Feed
## Executive GIST (TL;DR)
- The study reanalyzed a recent real-world benchmark that applied 2,234 distinct harmonization approaches to 7,174 **B-cell lymphoma** bulk transcriptomes, combining 31 methods, 14 batch-removal strategies, three imputations and two post-removal settings. - Investigators assigned 87 polarity-adjusted scoring metrics (from 344 registered) into five algorithmic classes—global, local, distributional, structural and composite—and added three biology-facing metric classes: within-cohort marker correlation (L), cross-batch rank agreement of a marker panel (M), and leave-one-batch-out prediction (N). - Each metric class produced a top 112 approaches (top 5%), and the team measured overlaps between these top sets using the Jaccard index to quantify agreement across metric classes. - Methods with strong distributional enforcement, such as quantile normalization, aligned batch distributions but scored poorly on local and global metrics, indicating trade-offs between distributional matching and biological signal retention. - The leave-one-batch-out class (N)—a cross-batch predictive test of lymphoma subtypes and normal B cells—favored the **ComBat** family, which occupied 84 of 112 top positions in this class. - The cross-batch rank panel class (M) selected only two methods, 10_mnn and 12_scanorama, which the authors noted produced correlation patterns interpreted as biological corruption. - Overlap among best approaches varied: local and composite metrics had the largest intersection (71 shared approaches); other pairwise overlaps (L/M, M/N) showed no shared approaches, and triple overlap was significantly below random expectation. - Authors conclude that the optimal harmonization method depends heavily on the chosen quality metrics and that metric classes trade off against one another; they provide a six-step recipe for selecting metrics, thresholds and cross-batch tests. - The source did not report specific parameter settings, per-method hyperparameters, or detailed performance numbers beyond the aggregate counts and overlap descriptions reported above.
## Clinical Analysis & Structured Key Points
2,234 attempts to harmonize a multi-platform B-cell lymphoma bulk transcriptomic dataset: ComBat is superior by cross-batch prediction quality | bioRxiv Skip to main content New Results 2,234 attempts to harmonize a multi-platform B-cell lymphoma bulk transcriptomic dataset: ComBat is superior by cross-batch prediction quality View ORCID Profile Daniil Nikitin , View ORCID Profile Arsen Arakelyan doi: https://doi.org/10.64898/2026.10.08.757652 Daniil Nikitin Institute of Molecular Biology, National Academy of Science of the Republic of Armenia, Yerevan, 0014, Armenia Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Daniil Nikitin For correspondence: danya.nikitin.orel{at}gmail.com Arsen Arakelyan Institute of Molecular Biology, National Academy of Science of the Republic of Armenia, Yerevan, 0014, Armenia Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Arsen Arakelyan Abstract Info/History Metrics Supplementary material Data/Code Preview PDF Abstract Background. Cross-platform transcriptomic harmonization benchmarks rely on individual success metric each: correlation with a reference profile, classifier accuracy, or a composite imported from single-cell integration. We found no published study that computes multiple orthogonal metric classes over one set of harmonization approaches and measures overlapping between the best approaches by these metrics. Methods. We used the 2,234 approaches of the recent real-world benchmark on 7,174 B-cell lymphoma transcriptomes. They combine 31 harmonization methods, 14 batch-removal strategies, three imputations and two post-removal settings. We assigned its 87 polarity-adjusted scoring metrics, out of 344 registered, to five classes (global, local, distributional, structural and composite) and added three biology-facing classes: within-cohort marker correlation (L), cross-batch rank agreement of a marker panel (M), and leave-one-batch-out (LOBO) prediction of lymphoma subtypes and normal B cells (N). Each class resulted in top 112 approaches (5%), and we measured overlap by the Jaccard index. Results. Harmonization methods of high harshness such as quantile normalization matched different batches to the same shape of distribution but were inferior by local and global metrics. Local and composite metrics resulted in the largest overlap of best approaches (71 shared). Classes L and M, M and N shared no approach, whereas L and N shared one approach without harmonization under the rare-batches-removed strategy, and all three overlaps were significantly below the random expectation. Class N resulted in the ComBat family in 84 of its 112 places. Class M elected only 10_mnn and 12_scanorama, methods that completely corrupted biology by correlation. Conclusions. Best transcriptomic harmonization approach is highly contingent on the quality metrics used, with metric classes trade-offing between each other. We provide a six-step recipe for metric classes, their thresholds and the cross-batch tests for selecting the best approaches. Competing Interest Statement Arsen Arakelyan directs the institute given as affiliation 1, where Daniil Nikitin is a PhD candidate, and this article forms part of that thesis. Footnotes https://github.com/Nikit357/FL_harmonization https://github.com/Nikit357/ComboBatch https://github.com/Nikit357/Shambhala2_fast https://zenodo.org/records/22737294 Funder Information Declared Committee of Higher Education and Sciences MESCS of Armenia , 21AG-1F021 Copyright The copyright holder for this preprint is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license . Back to top Previous Next Posted October 15, 2026. Download PDF Supplementary Material Data/Code Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following 2,234 attempts to harmonize a multi-platform B-cell lymphoma bulk transcriptomic dataset: ComBat is superior by cross-batch prediction quality Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share 2,234 attempts to harmonize a multi-platform B-cell lymphoma bulk transcriptomic dataset: ComBat is superior by cross-batch prediction quality Daniil Nikitin , Arsen Arakelyan bioRxiv 2026.10.08.757652; doi: https://doi.org/10.64898/2026.10.08.757652 Share This Article: Copy Citation Tools 2,234 attempts to harmonize a multi-platform B-cell lymphoma bulk transcriptomic dataset: ComBat is superior by cross-batch prediction quality Daniil Nikitin , Arsen Arakelyan bioRxiv 2026.10.08.757652; doi: https://doi.org/10.64898/2026.10.08.757652 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Areas All Articles Animal Behavior and Cognition (8141) Biochemistry (18976) Bioengineering (15123) Bioinformatics (45083) Biophysics (22894) Cancer Biology (20020) Cell Biology (27223) Clinical Trials (138) Developmental Biology (14102) Ecology (21329) Epidemiology (2067) Evolutionary Biology (25698) Genetics (16293) Genomics (23751) Immunology (18944) Microbiology (43053) Molecular Biology (18263) Neuroscience (94703) Paleontology (709) Pathology (3020) Pharmacology and Toxicology (5160) Physiology (8242) Plant Biology (16199) Scientific Communication and Education (2113) Synthetic Biology (4622) Systems Biology (10349) Zoology (2414)
## Related Clinical Research

- [Naive-like T cell proportion predicts durable response to axi-cel in lymphoma](https://medichelpline.com/clinical-feed/nature-1-determinants-of-response-to-car-t-cell-therapy-in-a-multicenter-multi-trial.md)
- [Ifupinostat phase 2 trial in relapsed/refractory DLBCL: study report and open questions](https://medichelpline.com/clinical-feed/pubmed-42893991.md) (DOI: 10.1038/s41392-026-02926-8)
- [No prognostic value of c-MET, HLA class II, or CD68 by IHC in early-stage classic Hodgkin lymphoma](https://medichelpline.com/clinical-feed/plos-one-14-immunohistochemistry-for-c-met-hla-class-ii-and-cd68-does-not-show-prognostic.md)
- [mRNA LNP co-delivery of autoactivated cGAS and membrane-tethered IL-12 induces systemic tumor cont](https://medichelpline.com/clinical-feed/biorxiv-1-combining-an-autoactivated-cgas-mutant-with-membrane-tethered-il-12-via-mrna.md)
- [Sintilimab plus R-CHOP vs DA-EPOCH-R in PMBCL — article content unavailable on source page](https://medichelpline.com/clinical-feed/frontiers-in-immunology-15-sintilimab-plus-r-chop-versus-da-epoch-r-as-first-line-therapy-for-primary.md)

## Navigation
- [← Back to Oncology Feed](https://medichelpline.com/clinical-feed/oncology.md)
- [← All Clinical Specialties](https://medichelpline.com/clinical-feed.md)
## Medical & Regulatory Disclaimer

> [!CAUTION]
> MedicHelpline content is structured for research, educational, and professional discovery purposes. It does not constitute individual medical advice, clinical diagnosis, or treatment recommendations.
> Always verify dosing, contraindications, and regulatory alerts against official product labeling and primary regulatory sources before clinical decision-making.