search

matsengrp / cft

Clonal family tree
5 stars 3 forks source link

Logprobs are incorrect? #114

Closed lauradoepker closed 7 years ago

lauradoepker commented 7 years ago

@psathyrella just explained the logprobs to me the other day and the data that are up online don't seem to match with that I learned. Clustering step 0 should have the highest probability (lowest absolute value negative number) and clustering steps 1-n should have progressively lower probabilities (larger absolute value negative numbers). This isn't what's happening online on stoat:5555. @metasoarous are the data being pulled in the right order? Can someone straighten me out here?

metasoarous commented 7 years ago

You're correct! My mistake. There is an indexing error in the code that puts that data together. I'll fix this.