Choose a pre-loaded organism (Salmonella SL1344 or B. thetaiotaomicron) from the dropdown, or select "Own files" to upload your own genome FASTA and GFF annotation.
For pre-loaded organisms a default sRNA is available (e.g. PinT for Salmonella, MasB for B. theta). Select "Own files" to upload your own sRNA FASTA.
Upload one or more CSV or Excel files containing experimental evidence.
Each file must include a gene identifier column (matching the GFF) and a column named
p_value
.
For pre-loaded organisms, default datasets are available and pre-selected.
After the pipeline completes, an interactive scatter plot and a downloadable results table are shown. Use the sidebar to change plot axes. Download results as TSV, Excel, or an interactive HTML report.
SPIRIT uses IntaRNA to predict sRNA–mRNA interactions based on minimum free energy (MFE) of hybridisation. Statistical significance is assessed by comparing real MFEs against a Gumbel distribution fitted to shuffled-sequence controls.
P-values from each evidence source (IntaRNA + experiment tables) are combined using Fisher's method
(optionally weighted). The test statistic is:
-2 * sum(w_i * log(p_i))
.
As an alternative, Stouffer's Z-method converts each p-value to a Z-score and combines them:
Z = sum(w_i * qnorm(1 - p_i)) / sqrt(sum(w_i^2))
.
Both Fisher and Stouffer combined p-values are adjusted for multiple testing using the Benjamini–Hochberg procedure.
https://github.com/jakobjung/spirit_server
