The past two years have seen the rapid development of new recognition
methods for protein structure prediction. These algorithms 'thread' th
e sequence of one protein through the known structure of another, look
ing for an alignment that corresponds to an energetically favorable mo
del structure. Because they are based on energy calculation, rather th
an evolutionary distance, these methods extend the possibility of stru
cture prediction by comparative modeling to a larger class of new sequ
ences, where similarity to known structures is recognizable by no othe
r means. The strength of the evidence they offer should be judged by o
bjective statistical tests, however, so as to rule out the possibility
that favorable scores arise from chance factors such as similarity of
length, composition, or the consideration of a large number of altern
ative alignments. Calculation of objective p-values by analytical mean
s is not yet possible, but it would appear that approximate values may
be obtained by simulation, as they are in gapped, global sequence ali
gnment. We propose that the results of threading experiments should in
clude Z-scores relative to the composition-corrected score distributio
n obtained for shuffled and optimally aligned sequences.