spsSIGNATURE has been developed to correct against such systemic biases and optimally deal with overfitting, predicting robust risk signatures to enable reproducible cohort stratifications. The pipeline can tackle multiple clinical outcome types, from time-to-event and ordinal class outcomes to real-valued clinical scales. spsSIGNATURE offers robust analysis for both small datasets with very few samples and those high-dimensional datasets containing many covariates. These advantages offer the chance of reliable cohort stratification and outcome prediction even with meagre datasets, for example, even for rare diseases or under-powered clinical trials; they also provide the opportunity to introduce additional variables into your analyses such as, for example, covariate-covariate interaction terms to gain a more nuanced insight into how different disease risk factors operate together. 
 

The uniqueness of our spsSIGNATURE algorithms lies in the replica method, setting them apart from the conventional regularisation techniques found in other statistical software. No data is sacrificed for hyperparameter optimisation-they're either unnecessary (in bias removal mode) or computed analytically (in regularisation mode).

Other features include

• ​Handling of informative covariate missingness.

• Inclusion of covariate interactions if required.

• Shadow analysis (analysis replication with outcome-randomised data, to exclude false positive inferences).

• Generation of simulated clinical data.

• Scripting.

• Validation of predictions on further data sets.

• Fully automated report generation.