{"apiVersion":"1","version":"sha256:faade27e0b75b2aca6311142ff65347b4dacc153e8d28cbdf0e8a237319331f9","data":{"id":"S7526","title":"Graph Fourier transform for spatial omics representation and analyses of complex organs","authors":["Yuzhou Chang","Jixin Liu","Yi Jiang","Anjun Ma","Yao Yu Yeo","Qi Guo","Megan McNutt","Jordan E Krull","Scott J Rodig","Dan H Barouch","Garry P Nolan","Dong Xu","Sizun Jiang","Zihai Li","Bingqiang Liu","Qin Ma"],"publishedDate":"2024-08-29","format":"Paper","url":"https://www.ebi.ac.uk/europepmc/webservices/rest/PMC11362340/fullTextXML","summary":"SpaGFT represents spatial data using graph topology and low-frequency assumptions. Benchmarks vary by dataset, and some methods perform better in particular examples. Selected marker references, repeated tissue regions and alternate analysis methods do not independently establish biological causation or patient replication.","subjects":["Scientific background","Spatial data analysis","Benchmark limitations"],"people":["Garry Nolan"],"review":{"label":"Main text and selected supporting materials read","scope":"The supplied main text, references and recorded static apparatus were read, with the previously recorded original-layout and fine-glyph limits. All 19 received MOESM4 worksheets and all 11 received MOESM6 worksheets are now complete within their recorded cell scopes. The final benchmark worksheet covers A1:D20002: 80,008 values and types across 101 actual untruncated reading returns, independently matched to the native workbook by root with no gaps or overlap. Its title and column header were read once. CAMPA and modified CAMPA each contribute 10,000 dependent model/parameter/iteration records, covering k_1–k_100 with 100 iterations each. These are benchmark repetitions, not independent subjects. Four required corrections and the excluded oversized initial return remain preserved without duplicate credit. Completing the received worksheets does not establish raw repository access, executable replication, calibrated performance or whole original-source completion.","gaps":["No supplied MOESM4 or MOESM6 worksheet cells remain unread. Original-layout main PDF, unreadable mathematical glyphs, raw repositories and executable analyses remain unresolved. The native kBET acceptance-versus-rejection direction and figure-label concordance conflicts remain; worksheet completion does not resolve them.","Low-frequency assumptions may suppress meaningful signals; selected markers are not exhaustive ground truth. Reviewer objections and author replies remain separate. Printed formula, direction, resolution, sample-count and reporting-summary inconsistencies are unresolved rather than proven code defects. Nolan’s role includes experiments, interpretation and shared writing, without stated algorithm coding or correspondence authorship. No UAP cohort, clinical recommendation or independently reproduced result is established.","Published tables retain missing-result markers, absent or incomplete settings, redundant metrics, significant Blank-labelled channels and unresolved fraction denominators. Tuning maxima and precision/recall tradeoffs are not independent scientific replication. Figure 2c retains 102 literal NA metrics, eight blank headers, a five-tool description versus six named competitors and repeated SPARK-X keys with unknown distinguishing settings.","The marker table contains 848 distinct literal mouse-symbol strings versus the article’s stated 849 candidates. Case variants and mouse/human mappings remain unnormalized; repeated literature attributions are not independent biological samples or newly read underlying studies.","The six-method table compares a selected union with incomplete coverage of the earlier selected-gene set. Unselected zeros do not imply absent expression. Repeated experiment vectors and native rank gaps remain qualified; expressing-spot counts do not count donors. Native decimal lexemes are preserved separately from equivalent binary scalar text. Statistical results were not independently recalculated.","The assignment description says B.follicle while native labels say B.zone. Numeric zero means no assignment; physical coordinate units remain unknown. Agreement with the same-source figure counts is not independent biological validation or a donor count.","The composition table supplies inferred proportions, not directly observed cell counts. Ordered spots are not a representative random sample or donor census. Its description conflicts between human lymph node and a mouse-sample label. Native decimal text and equivalent binary scalar precision remain separately preserved; a truncated reading return earned no credit and was repaired with smaller complete returns.","Source-data sheets retain mismatched figure labels and description order, a five-tool description versus six named competitors, repeated keys with unknown distinguishing parameters, and method-specific counterexamples. NA strings, negative values and merged blank cells remain distinct. Formula sign conventions differ across simulations; source-stored caches are not fresh calculations. Tuning and repeated dataset labels do not establish donor counts or independent replication.","The selected-gene table does not supply the entire tested-gene universe or independently validate significance, donor representation, gene categories or causal biology. Spatial spot counts are not patient counts; categorical cluster assignments do not establish diagnoses.","Two differently named macrophage columns have identical decoded values and original decimal strings across all 4,035 read spots. Their cause is unknown and they are not independent estimates. Numeric labels and coordinate zeros differ from missingness; scientific-notation strings retain their source types. Zone categories and enrichment maxima can disagree; derived scores do not establish calibrated probabilities, representative donors or diagnoses."]},"familyId":"10.1038/s41467-024-51590-5","citationUrl":"https://bioparity.io/sources/S7526","citation":"Yuzhou Chang, Jixin Liu, Yi Jiang, Anjun Ma, Yao Yu Yeo, Qi Guo, Megan McNutt, Jordan E Krull, Scott J Rodig, Dan H Barouch, Garry P Nolan, Dong Xu, Sizun Jiang, Zihai Li, Bingqiang Liu, Qin Ma. Graph Fourier transform for spatial omics representation and analyses of complex organs. August 29, 2024. BioParity S7526. https://bioparity.io/sources/S7526"}}