hoffmangroup / segway

Application for semi-automated genomic annotation.
http://segway.hoffmanlab.org/
GNU General Public License v2.0
13 stars 7 forks source link

Fatal fdiffact1 error when using --semisupervised and --validation-coords together #120

Open EricR86 opened 6 years ago

EricR86 commented 6 years ago

Original report (BitBucket issue) by Francis Nguyen (Bitbucket: Francis Nguyen).


When running segway train with --semisupervised and --validation-coords=<x> together, segway will submit jobs, but then fail with the following error:

ERROR: observation file 1 needs an -fdiffact1 that truncates (ts, te) or expands (rf, rl, se) because the observation files have different lengths for segment 0. All files must be truncated to 5 frames or expanded to 10 frames.

ERROR: observation file 1 needs an -fdiffact1 that truncates (ts, te) or expands (rf, rl, se) because the observation files have different lengths for segment 0. All files must be truncated to 5 frames or expanded to 10 frames.

This error can be replicated by taking the simplesemisupervised example in test/semisupervised, adding the line --validation-coords="../validate-coords.bed" \ after line 32 of run.sh, and creating a validate-coords.bed file in that directory, with contents:

chr1    1000    1100
EricR86 commented 6 years ago

Original comment by Eric Roberts (Bitbucket: ericr86, GitHub: ericr86).


A probable cause is that validation coords are not storing the supervised observations. I believe -of1 does refer to the int observations.