The service requires one to upload counts of samples in one table (“Counts Table”), and optionally, a table delineating sample groupings (“Labels Table”).
The Counts Table must have the following characteristics:
- Whole transcriptome gene expression (non-normalized) counts generated by sequencing on an Illumina sequencing machine and aligning to a human genome.
- A minimum of 10M total counts per sample and a maximum of 100 samples.
- A comma delimited (CSV) table, whose rows are genes and whose columns are samples.
- Ensembl gene ids.
Although the service will work with non-normalized counts data from any bioinformatics pipeline, we strongly recommend:
- aligning to the hg38 reference.
- using STAR to align and htseq to generate counts.
The Labels Table must be a comma delimited (CSV) table with two columns: one for sample name and the other for the sample’s group label.
An example label file is found here: https://tcsp.cofactorgenomics.com/pass_PD1.labels.csv for reference.
We strongly recommend:
- 4 or fewer unique group labels
- Group labels with less than 10 characters