Useful information about the output of PEPPI

This forum is shown on the index page along with all topics.

Moderator: robpearc

XiZhang
Posts: 21
Joined: Tue May 04, 2021 5:06 pm

Useful information about the output of PEPPI

Post by XiZhang »

Dear users,

Dr. Eric Bell, first author of the PEPPI, wrote the important information about the output of PEPPI. I post them here so that you could have a better understanding of various information in the PEPPI result page.

##############################################################################

 log(LR): This is the log likelihood ratio that your pair of sequences is interacting (log(LR) > 0 more likely to interact than not, log(LR) < 0 more likely to not interact)

SPRING: This is a dimeric threading program which identifies known interacting structure templates from an internal database and builds a rough model of your interaction.  Explanations of each column are provided underneath the table.

BLAST results: These results are derived by BLAST searching your sequences through a database of sequences which participate in interactions according to high throughput experimental data.  The total sequence ID is just a mean of the sequence IDs of the two chains individually.

STRING results: The STRING database is a functional, intra-species interaction database, and its functional association data can be helpful to doing the species agnostic physical interaction classification that PEPPI is built for.  If the interaction you're querying is found in STRING, the functional association features provided by STRING will show up here.  More often than not though, the interaction will not be found in STRING. CT: This is a neural network classifier, therefore the only output is the predicted probability of interaction according to the NN model.

SPRINGNEG: This is the same as SPRING, but instead queries a database of functionally associated but not interacting protein chains.  A high score in this module actually implies a LOWER likelihood of interaction, as the score implies likelihood of false positives classification.