Paper · S7679
Integration of spatial and single-cell data across modalities with weakly linked features
This Nolan-coauthored study introduces MaxFuse, a computational method for aligning cell measurements across weakly linked data types. Benchmarks report gains on selected datasets; inferred spatial gene activity is distinguished from measurements actually made in each cell. Reused cell batches and tuned comparisons limit broader conclusions. The article explicitly attributes FDA Medical Countermeasures Initiative contracts and NIH support to Nolan, without reporting a secret approach or classified UAP programme.
Read the paper at ebi.ac.ukWhat we checked
Read all 94,976 normalized supplied XML characters in ten contiguous parts, including the abstract, body, methods, five main and ten extended-data figure captions, 10-author metadata, funding/contribution/interest declarations and 66 reference entries. Native graphics, PDF, supplementary figures 1–13, tables 1–5, reporting summary, datasets, software and cited originals remain unreviewed.
Complete supplied XML text; original figures and supplements unreviewed
How we check sources
Not checked yet
- All supplied XML text, methods, 15 figure captions and 66 reference entries were read. Main and extended-data figure pixels, native PDF, supplementary figures 1–13 and tables 1–5, reporting summary, code, datasets and cited original bodies remain unreviewed. Flattened XML equations do not establish full native mathematical layout review or software reproduction.
- The article pub-date is September 7, 2023; issue metadata is volume 42, issue 7, 2024, with a December 13, 2024 repository/collection date. Its January 16, 2023 preprint has a slightly different title. These dates and versions are distinct, not three independently read studies.
- Benchmark gains are metric- and dataset-specific. Five random cell batches do not establish five independent donor cohorts. Reduced feature panels were ranked using reference cell-type labels. Strong-linkage accuracy was comparable to scGLUE; historical hardware-dependent runtime comparisons are not current performance guarantees.
- Spatial RNA and chromatin profiles are transferred or imputed from matched cells, including cells from different donors, rather than simultaneously measured with protein in every cell. Selected expression-pattern checks and two reported experimental validations support these examples without proving every match or a clinical outcome.
- The study attributes FDA Medical Countermeasures Initiative contracts HHSF223201610018C and 75F40120C00176 to Nolan, plus shared NIH award 3U54HG010426. Original contracts, grant files and any private approaches remain unreviewed. Disclosed funding and the authors’ views disclaimer do not establish classified access or a UAP programme.
- Nolan is one of 10 authors and one of three joint supervisors; algorithm implementation is credited to other named coauthors. His pharmaceutical research grants and Akoya equity/advisory role are disclosed. Author statements are not an independent financial audit. Data/code links were not downloaded or executed; no private patient records or experimental procedures were collected.
Cite this record
Shuxiao Chen, Bokai Zhu, Sijia Huang, John W Hickey, Kevin Z Lin, Michael Snyder, William J Greenleaf, Garry P Nolan, Nancy R Zhang, Zongming Ma. Integration of spatial and single-cell data across modalities with weakly linked features. September 7, 2023. BioParity S7679. https://bioparity.io/sources/S7679Record data (JSON)