FAQ¶
Data upload¶
What file formats are supported? CSV only. See Data Formats for structure requirements.
My upload fails — what should I check?
The most common causes are: non-numeric values in the matrix (check for text, units, or % symbols), sample IDs that don't match between the matrix and metadata, and files exported from Excel with a BOM character. See Data Formats — Common Pitfalls.
Can I upload multiple files for one dataset? Yes. Each dataset can have a matrix, sample metadata, and feature metadata file. Upload them as separate files and assign each a role during configuration.
How large can my files be? There is no hard limit, but matrices larger than ~100,000 features may be slow during preprocessing. Consider filtering to the most variable features before upload if performance is a concern.
Preprocessing¶
Should I log-transform my data before uploading? If your data is already log-transformed (e.g. log2 normalised expression, log-scale LC-MS from DepMap), tick "already log-transformed" during configuration and GrAndMA will skip the log step. If your data is in raw intensity units, leave it unticked and select log2 during preprocessing.
What imputation method should I use? For metabolomics with <20% missingness, KNN (k=5) works well. For data with structural missingness (features absent from entire groups), consider minimum-value imputation or none.
Can I re-preprocess a dataset? Yes. Go to the preprocessing page for that dataset and run again with different settings. Previous runs are preserved and you can switch between them.
Analysis¶
What statistical test is used for differential abundance? By default, a Welch t-test (two groups) or one-way ANOVA (three or more groups) with Benjamini–Hochberg FDR correction. Wilcoxon rank-sum and Kruskal–Wallis non-parametric alternatives are also available.
How do I analyse longitudinal data? Use the Stats module and select the mixed-effects or repeated-measures model. You will need a Subject ID column and a TimePoint column in your sample metadata.
Can I analyse multiple datasets together? Yes, if they share sample IDs. Use the Multi-Omics module to combine datasets from the same project. See Multi-Omics.
Results¶
How do I download my results? Each analysis result has a download button for CSV export. Plots can be downloaded as SVG or PNG from the toolbar in the top-right of each figure.
Results disappeared after I came back — what happened? Analysis results are linked to your account and persist between sessions. If you are using a demo account, results may be cleared periodically. Log in with a personal account to retain results.
Issues and feedback¶
I found a bug. Where do I report it? Open an issue in this repository with steps to reproduce.
I have a feature request. Open a feature request.
I have a private or sensitive question. Use the contact form at insilijo.github.io.