Skip to content
Snippets Groups Projects
Commit 941006d4 authored by Studer Gabriel's avatar Studer Gabriel
Browse files

revert qsscoring.py/compare-structures to state of release 2.3.1 

This code is considered deprecated as of release 2.4.0. The new lDDT
implementation has already been integrated into
qsscoring.py/compare-structures. I prefer to leave this completely legacy.
parent 4667f3a8
No related branches found
No related tags found
No related merge requests found
Expand all Hide whitespace changes
Inline Side-by-side
actions/ost-compare-structures
+ 112
− 108
View file @ 941006d4
......@@ -67,10 +67,12 @@ import ost
from ost.io import (LoadPDB, LoadMMCIF, SavePDB, MMCifInfoBioUnit, MMCifInfo,
MMCifInfoTransOp, ReadStereoChemicalPropsFile, profiles)
from ost import PushVerbosityLevel
from ost.mol.alg import (qsscoring, lddt, Molck, MolckSettings, lDDTSettings,
CheckStructure)
from ost.mol.alg import (qsscoring, Molck, MolckSettings, lDDTSettings,
CheckStructure, ResidueNamesMatch)
from ost.conop import (CompoundLib, SetDefaultLib, GetDefaultLib,
RuleBasedProcessor)
from ost.seq.alg.renumber import Renumber
def _GetDefaultShareFilePath(filename):
"""Look for filename in working directory and OST shared data path.
......@@ -518,19 +520,10 @@ def _RevertChainNames(ent):
def _CheckConsistency(alignments, log_error):
is_cons = True
for alignment in alignments:
for col in alignment:
r1 = col.GetResidue(0)
r2 = col.GetResidue(1)
if r1.IsValid() and r2.IsValid():
if r1.GetName() != r2.GetName():
is_cons = False
msg = f"Name mismatch for model residue {r2}: in the "
msg += f"reference structure(s) is {r1.GetName()} ({r1}), "
msg += f"in the model {r2.GetName()}"
if log_error:
ost.LogError(msg)
else:
ost.LogWarning(msg)
ref_chain = Renumber(alignment.GetSequence(0)).CreateFullView()
mdl_chain = Renumber(alignment.GetSequence(1)).CreateFullView()
new_is_cons = ResidueNamesMatch(mdl_chain, ref_chain, log_error)
is_cons = is_cons and new_is_cons
return is_cons
......@@ -844,24 +837,25 @@ def _Main():
"best_score": 0.0}
# Calculate lDDT
if opts.lddt:
lddt_settings = lDDTSettings(radius=opts.inclusion_radius,
sequence_separation=opts.sequence_separation, label="lddt")
ost.LogInfo("-" * 80)
ost.LogInfo("Computing lDDT scores")
ost.LogInfo("lDDT settings: ")
ost.LogInfo(str(lddt_settings).rstrip())
ost.LogInfo("===")
lddt_results = {
"single_chain_lddt": list()
}
if skip_score:
for aln in qs_scorer.alignments:
reference_chain = aln.GetSequence(0).GetName()
model_chain = aln.GetSequence(1).GetName()
lddt_settings = lDDTSettings(
radius=opts.inclusion_radius,
sequence_separation=opts.sequence_separation,
label="lddt")
ost.LogInfo("lDDT settings: ")
ost.LogInfo(str(lddt_settings).rstrip())
ost.LogInfo("===")
oligo_lddt_scorer = qs_scorer.GetOligoLDDTScorer(lddt_settings)
for mapped_lddt_scorer in oligo_lddt_scorer.mapped_lddt_scorers:
# Get data
lddt_scorer = mapped_lddt_scorer.lddt_scorer
model_chain = mapped_lddt_scorer.model_chain_name
reference_chain = mapped_lddt_scorer.reference_chain_name
if skip_score:
ost.LogInfo(
" --> Skipping single chain lDDT because "
"consistency check failed")
......@@ -873,97 +867,107 @@ def _Main():
"global_score": 0.0,
"conserved_contacts": 0.0,
"total_contacts": 0.0})
else:
try:
ost.LogInfo((" --> Computing lDDT between model "
"chain %s and reference chain %s") % (
model_chain,
reference_chain))
ost.LogInfo("Global LDDT score: %.4f" %
lddt_scorer.global_score)
ost.LogInfo(
"(%i conserved distances out of %i checked, over "
"%i thresholds)" % (lddt_scorer.conserved_contacts,
lddt_scorer.total_contacts,
len(lddt_settings.cutoffs)))
sc_lddt_scores = {
"status": "SUCCESS",
"error": "",
"model_chain": model_chain,
"reference_chain": reference_chain,
"global_score": lddt_scorer.global_score,
"conserved_contacts":
lddt_scorer.conserved_contacts,
"total_contacts": lddt_scorer.total_contacts}
if opts.save_per_residue_scores:
per_residue_sc = \
mapped_lddt_scorer.GetPerResidueScores()
ost.LogInfo("Per residue local lDDT (reference):")
ost.LogInfo("Chain\tResidue Number\tResidue Name"
"\tlDDT\tConserved Contacts\tTotal "
"Contacts")
for prs_scores in per_residue_sc:
ost.LogInfo("%s\t%i\t%s\t%.4f\t%i\t%i" % (
reference_chain,
prs_scores["residue_number"],
prs_scores["residue_name"],
prs_scores["lddt"],
prs_scores["conserved_contacts"],
prs_scores["total_contacts"]))
sc_lddt_scores["per_residue_scores"] = \
per_residue_sc
lddt_results["single_chain_lddt"].append(
sc_lddt_scores)
except Exception as ex:
ost.LogError('Single chain lDDT failed:', str(ex))
lddt_results["single_chain_lddt"].append({
"status": "FAILURE",
"error": str(ex),
"model_chain": model_chain,
"reference_chain": reference_chain,
"global_score": 0.0,
"conserved_contacts": 0.0,
"total_contacts": 0.0})
# perform oligo lddt scoring
if skip_score:
ost.LogInfo(
" --> Skipping oligomeric lDDT because consistency "
"check failed")
lddt_results["oligo_lddt"] = {
"status": "FAILURE",
"error": "Consistency check failed.",
"global_score": 0.0,
"conserved_contacts": 0,
"total_contacts": 0}
"global_score": 0.0}
else:
lddt_scorer = qs_scorer.GetOligoLDDTScorer(lddt_settings,
penalize_extra_chains=True)
# do single chain lDDT
scores = lddt_scorer.sc_lddt
per_res = lddt_scorer.sc_lddt_per_res
tot = lddt_scorer.sc_lddt_tot
cons = lddt_scorer.sc_lddt_cons
alns = qs_scorer.alignments
for idx in range(len(scores)):
ref_chain = alns[idx].GetSequence(0).GetName()
model_chain = alns[idx].GetSequence(1).GetName()
ost.LogInfo((" --> Computing lDDT between model "
"chain %s and reference chain %s") % (
model_chain,
ref_chain))
ost.LogInfo("Global LDDT score: %.4f" %scores[idx])
ost.LogInfo("(%i conserved distances out of %i checked, over "
"%i thresholds)" % (cons[idx], tot[idx], 4))
sc_lddt_scores = {
try:
ost.LogInfo(' --> Computing oligomeric lDDT score')
lddt_results["oligo_lddt"] = {
"status": "SUCCESS",
"error": "",
"model_chain": model_chain,
"reference_chain": ref_chain,
"global_score": scores[idx],
"conserved_contacts": cons[idx],
"total_contacts": tot[idx]}
if opts.save_per_residue_scores:
tmp = [x for x in per_res if x["chain"]==ref_chain]
ost.LogInfo("Per residue local lDDT (reference):")
ost.LogInfo("Chain\tResidue Number\tResidue Name"
"\tlDDT\tConserved Contacts\tTotal "
"Contacts")
for x in tmp:
ost.LogInfo("%s\t%i\t%s\t%.4f\t%i\t%i" % (
x["chain"],
x["residue_number"],
x["residue_name"],
x["lddt"],
x["conserved_contacts"],
x["total_contacts"]))
sc_lddt_scores["per_residue_scores"] = tmp
lddt_results["single_chain_lddt"].append(
sc_lddt_scores)
# perform oligo lddt scoring
ost.LogInfo(' --> Computing oligomeric lDDT score')
lddt_results["oligo_lddt"] = {
"status": "SUCCESS",
"error": "",
"global_score": lddt_scorer.oligo_lddt,
"conserved_contacts": lddt_scorer.oligo_lddt_cons,
"total_contacts": lddt_scorer.oligo_lddt_tot}
"global_score": oligo_lddt_scorer.oligo_lddt}
ost.LogInfo(
"Oligo lDDT score: %.4f" %
oligo_lddt_scorer.oligo_lddt)
except Exception as ex:
ost.LogError('Oligo lDDT failed:', str(ex))
lddt_results["oligo_lddt"] = {
"status": "FAILURE",
"error": str(ex),
"global_score": 0.0}
if skip_score:
ost.LogInfo(
"Oligo lDDT score: %.4f" % lddt_scorer.oligo_lddt)
if opts.save_per_residue_scores:
tmp = lddt_scorer.oligo_lddt_per_res
ost.LogInfo("Per residue local oligo lDDT (reference):")
ost.LogInfo("Chain\tResidue Number\tResidue Name"
"\tlDDT\tConserved Contacts\tTotal "
"Contacts")
for x in tmp:
ost.LogInfo("%s\t%i\t%s\t%.4f\t%i\t%i" % (
x["chain"],
x["residue_number"],
x["residue_name"],
x["lddt"],
x["conserved_contacts"],
x["total_contacts"]))
lddt_results["oligo_lddt"]["per_residue_scores"] = tmp
" --> Skipping weighted lDDT because consistency "
"check failed")
lddt_results["weighted_lddt"] = {
"status": "FAILURE",
"error": "Consistency check failed.",
"global_score": 0.0}
else:
try:
ost.LogInfo(' --> Computing weighted lDDT score')
lddt_results["weighted_lddt"] = {
"status": "SUCCESS",
"error": "",
"global_score": oligo_lddt_scorer.weighted_lddt}
ost.LogInfo(
"Weighted lDDT score: %.4f" %
oligo_lddt_scorer.weighted_lddt)
except Exception as ex:
ost.LogError('Weighted lDDT failed:', str(ex))
lddt_results["weighted_lddt"] = {
"status": "FAILURE",
"error": str(ex),
"global_score": 0.0}
reference_results["lddt"] = lddt_results
model_results[reference_name] = reference_results
if opts.dump_structures:
ost.LogInfo("-" * 80)
......
modules/mol/alg/pymod/qsscoring.py
+ 476
− 211
View file @ 941006d4
This diff is collapsed.
modules/mol/alg/pymod/wrap_mol_alg.cc
+ 2
− 13
View file @ 941006d4
......@@ -172,9 +172,6 @@ object lDDTSettingsInitWrapper(tuple args, dict kwargs){
label);
}
/*
lDDTScorer is commented out to not collide with the new lDDTScorer class
that lives in Python
object lDDTScorerInitWrapper(tuple args, dict kwargs){
object self = args[0];
......@@ -220,7 +217,7 @@ object lDDTScorerInitWrapper(tuple args, dict kwargs){
model,
settings);
}
*/
void clean_lddt_references_wrapper(const list& reference_list)
{
......@@ -268,10 +265,6 @@ list get_lddt_per_residue_stats_wrapper(mol::EntityView& model,
return local_scores_list;
}
/*
lDDTScorer is commented out to not collide with the new lDDTScorer class
that lives in Python
list get_local_scores_wrapper(mol::alg::lDDTScorer& scorer) {
std::vector<mol::alg::lDDTLocalScore> scores = scorer.GetLocalScores();
list local_scores_list;
......@@ -289,7 +282,7 @@ list get_references_wrapper(mol::alg::lDDTScorer& scorer) {
}
return local_references_list;
}
*/
void print_lddt_per_residue_stats_wrapper(list& scores, bool structural_checks, int cutoffs_size){
int scores_length = boost::python::extract<int>(scores.attr("__len__")());
......@@ -402,9 +395,6 @@ BOOST_PYTHON_MODULE(_ost_mol_alg)
.def_readwrite("conserved_dist", &mol::alg::lDDTLocalScore::conserved_dist)
.def_readwrite("total_dist", &mol::alg::lDDTLocalScore::total_dist);
/*
lDDTScorer is commented out to not collide with the new lDDTScorer class
that lives in Python
class_<mol::alg::lDDTScorer>("lDDTScorer", no_init)
.def("__init__", raw_function(lDDTScorerInitWrapper))
.def(init<std::vector<mol::EntityView>&, mol::EntityView&, mol::alg::lDDTSettings&>())
......@@ -416,7 +406,6 @@ BOOST_PYTHON_MODULE(_ost_mol_alg)
.def_readonly("model", &mol::alg::lDDTScorer::model_view)
.add_property("references", &get_references_wrapper)
.add_property("is_valid", &mol::alg::lDDTScorer::IsValid);
*/
class_<mol::alg::StereoChemicalProps>("StereoChemicalProps",
init<mol::alg::StereoChemicalParams&,
......
modules/mol/alg/tests/test_qsscoring.py
+ 4
− 3
View file @ 941006d4
......@@ -424,12 +424,13 @@ class TestQSscore(unittest.TestCase):
lddt_oligo_scorer2 = qs_scorer2.GetOligoLDDTScorer(lddt_settings, False)
self.assertAlmostEqual(qs_scorer2.global_score, 0.171, 2)
self.assertAlmostEqual(qs_scorer2.best_score, 1.00, 2)
self.assertAlmostEqual(lddt_oligo_scorer2.weighted_lddt, 0.4996, 2)
self.assertAlmostEqual(lddt_oligo_scorer2.weighted_lddt, 1.00, 2)
self.assertEqual(len(lddt_oligo_scorer2.sc_lddt), 2)
self.assertAlmostEqual(lddt_oligo_scorer2.sc_lddt[0], 1.00, 2)
self.assertAlmostEqual(lddt_oligo_scorer2.sc_lddt[1], 1.00, 2)
self.assertAlmostEqual(lddt_oligo_scorer2.oligo_lddt, 0.4496, 2)
# penalty only affects additional model chains, scores are thus the same
# without penalty we don't see extra chains
self.assertAlmostEqual(lddt_oligo_scorer2.oligo_lddt, 1.00, 2)
# with penalty we account for extra reference chains
lddt_oligo_scorer2_pen = qs_scorer2.GetOligoLDDTScorer(lddt_settings, True)
self.assertAlmostEqual(lddt_oligo_scorer2_pen.oligo_lddt, 0.4496, 2)
self.assertAlmostEqual(lddt_oligo_scorer2_pen.weighted_lddt, 0.50, 2)
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment