gene_set_camera
Read-only
Competitive gene set test that accounts for inter-gene correlation (camera; Wu & Smyth 2012), as limma 3.68.5 runs it: is a set more differentially expressed than the genes outside it, with the variance of its mean statistic inflated by the correlation between its genes instead of treating them as independent. From a count table it fits voom (filterByExpr, TMM, precision weights) and tests the design's coefficient per set (camera); from a ranked statistic it tests that statistic directly (cameraPR). Parametric or rank-based, fixed or per-set estimated correlation, directional or not. Every set is bit-identical to limma in R on filtered pasilla voom and on DESeq2's Wald statistics against every fly GO term.
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| name | No | the tested coefficient by name; empty is the design's last. Only applies when statistic is voom. | |
| _files | No | The project: record references shaped like _subject. Their records become library documents a tool names in sequences/parts/document; a trace or alignment is found by file name. | |
| counts | No | a gene-by-sample table of raw read counts, gene IDs in the first column: camera then runs on its voom fit (filterByExpr, TMM, voom weights). Give this or `ranked`. Only applies when statistic is voom. | |
| design | No | with `counts`: the design formula, the variable of interest last. Only applies when statistic is voom. | ~ condition |
| ranked | No | a gene-wise statistic, 'gene statistic' per line (e.g. a DE run's Wald or moderated t). Only applies when statistic is ranked. | |
| samples | No | with `counts`: one row per sample, its name first, then the design's variables. Only applies when statistic is voom. | |
| _subject | No | The record to act on: exactly one of sequence, accession (+source), sample (an id from list_samples) or file ({name, text} or {name, base64}); record_index picks one record of several. | |
| contrast | No | 'factor, numerator, denominator', the denominator being the factor's reference level. Only applies when statistic is voom. | |
| geneSets | Yes | the gene sets, as GMT text (MSigDB, GO or WikiPathways exports): one set per line, tab-separated ID, description, then member gene IDs. IDs must be the same kind as the genes'. | |
| minCount | No | filterByExpr: minimum count in the smallest group's worth of samples. Only applies when statistic is voom. | |
| useRanks | No | a correlation-adjusted Wilcoxon rank-sum test instead of the parametric t on z-scores; robust to a few extreme genes. | |
| fdrCutoff | No | a set is called when its BH FDR (its p-value when one set is tested) is at most this. | |
| reference | No | reference level per factor, 'variable=level' separated by commas; a factor not named takes its alphabetically first level. Only applies when statistic is voom. | |
| statistic | No | ranked: cameraPR on a supplied gene-wise statistic (`ranked`). voom: camera on a voom fit of `counts` and `samples`, each gene's own residuals and weights. Given only `counts`, voom is assumed. | ranked |
| allowNegCor | No | with an estimated correlation: let a negative estimate deflate the variance (limma floors it at 0). Only applies when statistic is voom. | |
| directional | No | test up or down against the rest (true), or only whether the set's genes change more in either direction (false, a rank test on |statistic|). | |
| filterByExpr | No | drop genes too lowly expressed to test (edgeR's filterByExpr) before voom. Only applies when statistic is voom. | |
| interGeneCor | No | the fixed inter-gene correlation, limma's default 0.01 (a value most real sets exceed slightly). | |
| minTotalCount | No | filterByExpr: minimum total count. Only applies when statistic is voom. | |
| estimateCorrelation | No | with `counts`: estimate each set's correlation from the residuals instead of fixing it; the test then uses min(residual df, genes - 2) degrees of freedom. Only applies when statistic is voom. |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
| terms | Yes | every set with a member among the tested genes, by p: genes in the set, correlation, direction, p and BH FDR | |
| summary | Yes | the mode (voom camera or cameraPR), genes tested, sets supplied and tested, the degrees of freedom, the prior df of the variance squeeze, the settings limma changed | |
| enriched | Yes | the sets at or under the FDR cutoff |