v1.5 Specificity Score Scaled to Dataset Size

This version changes MiST's specificity score that is used for the combined MiST score. With the observation that specificity can depend on dataset size, we calculate a scaled specificity score that makes the scaled score more comparable with the original HIV study.

MiST now automatically uses the scaled specificity score (specificity18, by default) in the combined MiST score. See docs/ScalingSpecificity.pdf for a discussion of the math behind this change. This means MiST thresholds and weights should be closer to those used in the original study, but also scores will be sensitive to MiST version. Don't change versions mid-study without changing your yml file!

You can disable the use of specificity score scaling in the combined MiST score by changing your config yml file to include:

  standardize_specificity : FALSE

See tests/small/mist_small_test.yml for an example config yml file.

We do not recommend disabling this in general, but it can be useful for consistency with previous versions. Regardless of the setting, the output will include columns for both the scaled and unscaled specificity; this setting changes which specificity score will be used in MiST score.

Original MiST with license

Very small changes from original release, now with License added. Release made for purposes of linking to publication.

stable version of the MiST pipeline

stable version of the MiST pipeline, reflecting the description in the current protocols in bioinformatics publication (Verschueren et al, 2014).