Merging Alignments
This code is adapted from the SEPP codebase, so you'll need to have a working installation of SEPP. It also requires you to have installed the bundled HMMER executables, although you can play with the config file as necessary. Change the path names to your own, since all of the uploaded ones are my own absolute paths.
The code is intended to help someone merge two hmmalign results. That is, given a backbone alignment on {S1, S2, S3, S4} and two query sequences x and y, if we build an HMM on {S1, S2} and another on {S3, S4}using HMMER, then alignxto{S1, S2}and aligny$ to {S3, S4}, we would like to be able to create a final alignment with {S1, S2, S3, S4, x, y}.
query_x.txt and query_y.txt show the two query sequences. I've also provided two scripts run_gap.sh and run_nogap.sh to illustrate how to generate the other transient input files necessary. Note that I have both s1_subset.txt, s2_subset.txt and a second pair s1_subset2.txt and s2_subset2.txt to illustrate the difference in outputs if you include gaps in the subset sequences passed to the SEPP code. In particular, you want to follow the nogaps pipeline in order to get a final alignment.
Test Queries
These test queries are generated on 1000M1, 1000M4. They are taken from the set of query sequences (i.e. all fragmentary sequences) that UPP finds as a first step in the UPP pipeline. These test queries are those that pass some threshold (defined in the file name) of fitting well with more than one HMM in the possible HMMs in the UPP pipeline.