Supplementary Figures for

D. Metzler (2002) Statistical alignment based on fragment insertion and deletion models
(submitted to Bioinformatics)

Software used in Sections 3.1 & 4.3: MCMCSALUT

Section 3.1 (Example: HVR-1 from human and orangutan)

Figure A:

The database alignment of a human (top sequence) and an orangutan HVR-1 sequence (bottom sequence) and the percentage of 1000 sampled alignments (with almost flat prior) that differ from it in each position.


Section 4.3 (Robustness of the sampling algorithm)

Figure B:

30 sequence pairs of length ~1000 were generated according to the FID (left) and the GID (right) model with s=0.5, lambda=0.1 and gamma=3. For each pair we sampled 50 alignment-mutation-parameters-combinations according to their posterior probability in the FID model. The point clouds show the overlay of the (lambda,gamma)-component of the sampling results.

Figure C:

The point clouds show the overlay of the (lambda,s)-component of the sampling results of the experiment described in the caption of Figure C.

Dirk Metzler
Last modified: Mon Oct 7 19:12:48 MSZ 2002