PlantGenIE — comparative layer · proof of concept

Live from the Neo4j graph via a JSON API. No database login — the browser only sees results.

Thresholds definitions
Gene list
One row per gene and genotype (Arabidopsis IDs give a row for each Arabidopsis genotype). DE: the call at the current thresholds; blank where the gene has no test at that timepoint. Co-expressolog partners: genes of the other genotype whose neighbourhoods overlap this gene's (co-expressolog P below the current threshold). The last columns count the other genotypes in which any member of the gene's orthogroup is differentially expressed at any timepoint. Inputs that match no gene of the study are listed above the table. Click a gene to open it.

Gene profile

Load a gene…

Co-expression network

The query gene (red) and all its neighbours in this experiment's co-expression network (highest reciprocal rank, HRR ≤ 100; the resource networks, not the pair-specific networks ComPlEx uses). Every neighbour is linked to the query gene; the strength of that link is shown by colour (darker blue = lower HRR = stronger); node size = the gene's number of co-expression links within the module. Lines are edges among the neighbours: width ∝ |Pearson r|, orange = negative correlation. By default each gene's 3 strongest are drawn, and in the force layout stronger edges pull genes closer. These genes are the module tested in Functional analysis. Hover to see a name, click a node to open it.

Response (differential expression)

All species

One line per genotype: the query gene and its counterpart in each genotype compared with it. The counterpart is the conserved-core clique member when the orthogroup is in the core, otherwise the strongest co-expressolog partner; the table says which rule chose each line, since they are not all "the orthologue". Filled point = differentially expressed at the current thresholds. Click a genotype in the legend to hide or show it.

The same gene in other studies

Every experiment, in any study in the graph, on this gene's annotation that has data for it. Genes are not mapped between annotation versions: where a study uses another version for this genotype, search there instead.

Candidate regulators

Transcription factors among the query gene's network neighbours (within the current HRR). A link is conserved in another genotype when the transcription factor's co-expressolog partner there is a network neighbour of the query gene's partner there (any partner pair counts). Ranked by the number of genotypes with a conserved link, then by HRR; candidates are drawn as diamonds in the network view. Candidates for follow-up, from co-expression; not evidence of regulation.

Functional analysis

Exploratory: GO biological-process enrichment of this gene's co-expression module (the network above), run with the paper's method and GO annotation — topGO on each experiment's expressed-gene background. elim (the default) tests specific terms first and removes their genes from parent terms; terms are listed at P < 0.001 by default, or at P < 0.005 or P < 0.01 (not adjusted for multiple testing; the help page explains the choice). classic, the test of the paper's tables, tests every term, BH-adjusted, significant at padj < 0.05. This is not a published result. Click a term to highlight its genes in the network; click a count to list the genes.
This gene's GO terms in PlantGenIE's hosted annotation (0) — not the paper's GO mapping

Orthologs across species

Each ortholog's response per experiment of the study (rows) and timepoint (columns): red induced, blue repressed (log₂ fold change); blank = no value. * = differentially expressed at that timepoint (paper's definition). ▶ marks the query gene. Click a gene above to open it.

Comparative co-expression (co-expressologs)

Co-expressolog partners of this gene as tested in this experiment (0, strongest effect first) — click one to compare neighbourhoods:
The query gene (red) and its partner (blue), each with its 40 strongest HRR co-expression neighbours. The evidence for the pair is the ComPlEx statistic above: for each gene, how many of its neighbours have an orthologue among the partner's neighbours, counted on ComPlEx's own networks for this comparison (|Pearson r|, mutual-rank normalised, top 3% of gene pairs, genes with an orthologue in the partner) — much larger neighbourhoods than those drawn here. The dashed green lines join orthologous genes that both appear in the two drawn neighbourhoods; they illustrate shared structure but are not the ComPlEx overlap. Click any node to open it.