Protein multiple sequence alignment benchmark.
There are three benchmarks: balifam100, balifam1000 and balifam10000.
These are Balibase v3 reference alignments with PFAM homologs added.
Directory structure under each benchmark is:
in/
Input sequences.
ref/
Reference alignments. Upper-case regions indicate conservative
regions that are intended for use in assessment. Lower-case regions
should not be used.
info/
Contains one file, ids.txt, which is a list of set identifiers used as filenames in ref/ and in/.
Accuracy of alignments made by an algorithm can be asssessed by qscore (https://github.com/rcedgar/qscore).
R.C. Edgar (2021), MUSCLE v5 enables improved estimates of phylogenetic tree confidence by ensemble boostrapping, biorxiv.