After running another bruco analysis (https://ldas-jobs.ligo-wa.caltech.edu/~pep.covas/bruco/1240824092/) I found coherence around these two lines with several channels related to SRM and PRM.

I found these lines in the coil driver monitors of PRM and SRM, and also in some sensors like LSC-REFL_A_RF45_I_ERR_DQ and LSC-REFL_B_RF9_Q_ERR_DQ. I traced these signals back to the controls signals of the loop LSC-SRCL_OUT and LSC-PRCL_OUT, which also show increased coherence around these frequencies.

I checked more channels with DTT, and apparently these lines are there for LSC signals, but not ASC signals. Jeff has suggested that some notches around these frequencies could be applied to the LSC loops in order to suppress these lines.