Documentation Getting Started Gene Prioritization
Browse Documentation
Introduction to PlantQTLdb Searching for Genes and Traits Navigating Gene-Centered Results Using Trait–Gene Evidence Gene Prioritization Connect ChatGPT, Codex and Scripts

Getting Started

Gene Prioritization

Find candidates from a genomic region or gene list, compare their evidence for a research goal, and save a report.

Choose your candidates

Select a species and check the displayed reference genome. Genomic Region accepts a variant position with left and right flanks, or a genomic interval. Coordinates are 1-based and inclusive; regions are limited to 10 Mb and 2,000 overlapping genes.

Gene List accepts up to 2,000 unique reference gene IDs, separated by spaces, commas or newlines. IDs are case-sensitive; gene symbols and transcript IDs are not mapped automatically. Paste IDs or use Upload file. TXT files contain one ID per line; CSV and TSV files require a single column named gene_id. Open File format for these requirements. Files use UTF-8 text and are limited to 256 KB; loading a file replaces the IDs in the input box. Unmatched IDs are listed separately and can be exported.

For a rice example, select Load rice example in either input route, then Find genes. The list example uses LOC_Os06g10880, LOC_Os06g10810 and LOC_Os06g10790 with the research goal drought response.

Set the research goal and run analysis

  1. Describe the goalEnter a keyword or short question in Research goal, including any relevant tissue, stage or treatment in the same field. Leave it blank to browse candidates; a goal is required for AI analysis. Context guides evidence relevance. Only explicit must-have requirements prevent ranking when unresolved; preferences do not exclude candidates.
  2. Retrieve candidatesFind genes returns the reference genes, catalog counts and literal text matches. Text matching ignores case but does not infer synonyms. Counts describe available records, not support for the research goal.
  3. Compare evidenceSelect Run analysis. If signed out, you will be taken to the sign-in page and returned here with your inputs and candidates retained. Select Run analysis again when ready. Assistant analysis uses the account's Assistant allowance and supports up to 80 candidates per run. It interprets the goal, selects relevant data tools and sources, and compares cited findings using criteria stated in the report. Each run may use different sources; it does not require every data layer. Simple expression or splicing/APA variability goals can use recorded profile statistics: the program calculates ranks and ties within one dataset, context, unit and replicate level. Event comparisons use the maximum SD among retrieved events and are descriptive, not differential or significance tests.

Read and filter the results

Order and Priority
Before analysis, Order is the table display order. After comparison, Priority is the ordinal review priority for this goal; ties share a level. Changing Sort by changes the display order without changing an assigned priority.
Not ranked and Not assessed
Not ranked means a gene was assessed but the retrieved evidence did not justify a priority. Not assessed means its evaluation is incomplete. Both remain in the candidate set; neither establishes low biological relevance.
Filter genes
Catalog records filters genes by available database records. Text matches filters annotation or literature keyword hits. Analysis results filters the completed assessment state. These filters change the displayed rows, not the research goal or analysis.
Evidence and Reason
Evidence identifies source types cited in the assessment; Reason explains the comparison. View details provides the full finding, next check and sources. Missing evidence does not establish biological absence, and a priority is not a causal probability.
Clarify the goal
A clarification prompt appears only when the interpretation is uncertain. Clarify goal returns to the Research goal field. Describe any corrections or must-have conditions there, then select Find genes and Run analysis. The report retains the interpreted goal and analysis criteria.

Save, return and rerun

Use View report to read the summary, comparison method and gene assessments, then Download HTML report to keep a copy. Evidence coverage at the end of the report shows retrieval status by gene and source, including searches that were incomplete or not performed. Returned records do not necessarily support the goal. Export all includes all retrieved genes despite table filters; Export selected includes your manually checked genes, including those currently hidden by filters.

Inputs, selections and reports are restored within the current browser tab when storage is available. Returning to the page or refreshing does not start another analysis. This page has no permanent report-history library; download important results before closing the tab, clearing browser storage or starting a new search.

Find genes starts a fresh candidate search and clears the current report. Changing inputs or input route also clears it. Run new analysis keeps the previous report available until a replacement succeeds; cancellation or failure retains that previous report. Partial reports state how many genes were assessed. Failed required searches leave affected genes unranked; other comparable genes can retain priorities. Supplementary search failures are disclosed without discarding usable evidence. Retry unfinished reuses saved evidence and completed assessments, then updates the comparison across all assessed candidates. Retry comparison repeats only that final step. Both use your Assistant allowance. Older reports without a saved recovery point require Run new analysis. Missing evidence is a valid assessment result, not an execution failure.