TIMER2.0 for analysis of tumor-infiltrating immune cells

TIMER2.0 for analysis of tumor-infiltrating immune cells Taiwen Li Jingxin Fu Zexian Zeng David Cohen Jing Li Qianming Chen Bo Li X. Shirley Liu

Tag layer of the DHSS Meta-Tagging Project. The article text is reproduced from the staged copy named in sourceDesc and is not altered; the tags are held separately in standOff so that overlapping and nested spans are preserved.

TIMER2.0 for analysis of tumor-infiltrating immune cells 2020 Nucleic Acids Research 10.1093/nar/gkaa407 li2020-timer2-0-for-analysis-of-tumor-inf li2020-timer2-0-for-analysis-of-tumor-inf.txt, the project's staged copy 9015a7659f57c19fb5243ecb40907dabd955f3d1 cc-by-nc medicine

the paper's own words

the tag layer; a sentence written by the tagger about the paper, always anchored to a quotation but not itself a quotation

0 spans located by searching the stored text for the quote's exact words, allowing any whitespace between them immune cell composition of tumours - a quantity that is inferred from bulk transcriptome rather than counted, by six different algorithms Quotation withheld because this record is not cleared for redistribution. how to make immune-infiltration estimates from public tumour data usable without committing to any one inference algorithm Quotation withheld because this record is not cleared for redistribution. a web platform that runs six estimation algorithms over uniformly pre-processed TCGA data and shows their outputs side by side Quotation withheld because this record is not cleared for redistribution. method Quotation withheld because this record is not cleared for redistribution. tumour immune infiltration, whose direct measurement is limited enough that the field works from computational inference instead Quotation withheld because this record is not cleared for redistribution. system-or-tool-building system-or-tool-building Quotation withheld because this record is not cleared for redistribution. gene-signature methods score cell-type-specific gene sets INDEPENDENTLY per cell type; deconvolution methods model the tissue's expression as a WEIGHTED SUM of the component cells' profiles. Two different mathematical objects behind one output name Quotation withheld because this record is not cleared for redistribution. THE QUANTITY IS NOT MEASURED. `Due to the limitations of direct measurement methods, computational algorithms are often used to infer immune cell composition` - and the transcriptome `does not immediately indicate` it Quotation withheld because this record is not cleared for redistribution. SIX ALGORITHMS, ONE OUTPUT NAME, AND ONLY TWO PRODUCE A QUANTITY ON A MEANINGFUL SCALE: `EPIC and quanTIseq have the advantage of directly generating scores interpreted as cell fractions`. The other four produce enrichment scores displayed beside them Quotation withheld because this record is not cleared for redistribution. each algorithm is named with its own failure: TIMER handles tissue specificity but covers only six cell types, xCell covers more but MAY FAIL TO DETECT SIGNALS FROM HOMOGENEOUS SAMPLES Quotation withheld because this record is not cleared for redistribution. the justification for running all six is a HOPE, in the authors' own word: `We hope that evaluating estimations from multiple algorithms might help the user gain more comprehensive and robust results` Quotation withheld because this record is not cleared for redistribution. and the cross-species use is caveated where it is offered: mouse estimations `should be interpreted with caution since human and the mouse tumors may have different immune properties` Quotation withheld because this record is not cleared for redistribution. all TCGA samples are pre-processed UNIFORMLY before the six algorithms run, so the estimates differ by algorithm and not by pipeline eliminates: disagreement between algorithms that is really disagreement between normalisations - which is the only way the side-by-side display could mean anything · outcome: held Quotation withheld because this record is not cleared for redistribution. whether the six estimates are on the same scale eliminates: nothing - EPIC and quanTIseq generate cell fractions and the rest do not, and the platform shows all six together without reconciling them · outcome: failed Quotation withheld because this record is not cleared for redistribution. whether running six algorithms actually yields more robust results eliminates: nothing that is demonstrated here - the claim is 'We hope ... might help', with no comparison against single-algorithm use reported · outcome: absent-declared-unreported Quotation withheld because this record is not cleared for redistribution. deconvolution deconvolution Quotation withheld because this record is not cleared for redistribution. cell fractions cell fractions Quotation withheld because this record is not cleared for redistribution. xCell xCell Quotation withheld because this record is not cleared for redistribution. TIMER TIMER Quotation withheld because this record is not cleared for redistribution. immune infiltration level - AS DELIVERED BY THIS PLATFORM, SIX DIFFERENT QUANTITIES UNDER ONE HEADING. Two are cell fractions, four are enrichment scores, and the platform's design is to show them together rather than to reconcile them. THE NAME IS THE ONLY THING THE SIX SHARE Quotation withheld because this record is not cleared for redistribution. estimation, AS OPPOSED TO MEASUREMENT - the paper's own framing throughout, and it is honest: direct measurement is limited, so the field infers. Every number this server is cited for is downstream of that sentence Quotation withheld because this record is not cleared for redistribution. robustness, AS THIS PLATFORM OPERATIONALISES IT: agreement across six algorithms, displayed for the user to judge. Not a statistic, not a test - A LAYOUT. And the authors say they HOPE it helps Quotation withheld because this record is not cleared for redistribution. critiques → reliance on any single immune-infiltration algorithm, since the two families have variable performance advantages across cell types and tumours Quotation withheld because this record is not cleared for redistribution. extends → the original TIMER web server, which used one algorithm where this uses six Quotation withheld because this record is not cleared for redistribution. THE QUANTITY IS INFERRED, NOT MEASURED: due to the limitations of direct measurement methods, immune cell composition is computed from bulk transcriptome THE QUANTITY IS INFERRED, NOT MEASURED: due to the limitations of direct measurement methods, immune cell composition is computed from bulk transcriptome Quotation withheld because this record is not cleared for redistribution. EPIC and quanTIseq directly generate scores interpreted as cell fractions EPIC and quanTIseq directly generate scores interpreted as cell fractions Quotation withheld because this record is not cleared for redistribution. each algorithm carries its own failure mode - TIMER covers only six cell types, xCell may fail to detect signals from homogeneous samples each algorithm carries its own failure mode - TIMER covers only six cell types, xCell may fail to detect signals from homogeneous samples Quotation withheld because this record is not cleared for redistribution. the reason for running all six is stated as a HOPE: `We hope that evaluating estimations from multiple algorithms might help the user gain more comprehensive and robust results` the reason for running all six is stated as a HOPE: `We hope that evaluating estimations from multiple algorithms might help the user gain more comprehensive and robust results` Quotation withheld because this record is not cleared for redistribution. mouse estimations SHOULD BE INTERPRETED WITH CAUTION because human and mouse tumours may have different immune properties - the caveat is placed where the feature is offered mouse estimations SHOULD BE INTERPRETED WITH CAUTION because human and mouse tumours may have different immune properties - the caveat is placed where the feature is offered Quotation withheld because this record is not cleared for redistribution. methods that model the gene expression of a tissue sample as the weighted sum of the expression profiles from the cells in the population mix Deconvolution Quotation withheld because this record is not cleared for redistribution. deconvolution deconvolution Quotation withheld because this record is not cleared for redistribution. tissue specificity tissue specificity Quotation withheld because this record is not cleared for redistribution.

The full text of this article is not redistributable, so it is not included here. The evidence sentences are withheld with it; the tag layer and its grounding status remain. The article itself is at https://doi.org/10.1093/nar/gkaa407.