Merge branch 'master' of https://github.com/steineggerlab/foldseek

steineggerlab · Nov 23, 2023 · 7180ed4 · 7180ed4
2 parents 28a4a7f + e00a3dc
commit 7180ed4
Show file tree

Hide file tree

Showing 29 changed files with 73 additions and 31 deletions.
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -16,7 +16,7 @@ set(FRAMEWORK_ONLY 1 CACHE INTERNAL "" FORCE)
 include(MMseqsSetupDerivedTarget)
 add_subdirectory(lib/mmseqs EXCLUDE_FROM_ALL)
 
-set(FOLDSEEK_FRAMEWORK_ONLY 0 CACHE INTERNAL "" FORCE)
+set(FOLDSEEK_FRAMEWORK_ONLY 0 CACHE BOOL "Framework mode (don't create foldseek executable)")
 if (FOLDSEEK_FRAMEWORK_ONLY)
     set(FRAMEWORK_ONLY 1 CACHE INTERNAL "" FORCE)
 endif()

diff --git a/README.md b/README.md
@@ -27,7 +27,7 @@ Foldseek enables fast and sensitive comparisons of large structure sets.
     - [Output](#output-cluster)
     - [Important Parameters](#important-cluster-parameters)
   - [Complexsearch](#complexsearch)
-    - [Output](#output-complexsearch)
+    - [Output](#complex-search-output)
 - [Main Modules](#main-modules)
 - [Examples](#examples)
 
@@ -218,14 +218,26 @@ MCAR...Q
 | --tmscore-threshold      | Alignment  | accept alignments with an alignment TMscore > thr                               |
 | --lddt-threshold      | Alignment  | accept alignments with an alignment LDDT score > thr                               |
 
-### Complexsearch 
-The `easy-complexsearch` module allows to search single or multiple query protein complexes, formatted in PDB/mmCIF format (flat or gzipped), against a target database, folder or single protein complexes. In default it outputs the alignment information as a [tab-separated file](#tab-separated-complex) but we support also [report](#report). <!-- or a HTML output. -->
 
-    foldseek easy-complexsearch example/1tim.pdb.gz example/8tim.pdb.gz aln tmpFolder
+### Complexsearch
+The `easy-complexsearch` module is a tool for searching single or multiple query protein complexes (PDB/mmCIF, flat or gzipped) against a  target database of protein complexes. It reports the similarity metrices of the complexes like TMscore.
 
-#### Output Complexsearch
+#### Using Complexsearch
+To pairwise compare complexes use `easy-complexsearch`, run the following command:
+```
+foldseek easy-complexsearch example/1tim.pdb.gz example/8tim.pdb.gz result tmpFolder
+```
+This command searches the specified protein complexe `1tim.pdb.gz` against 8tim.pdb.gz, producing alignment information.
+Foldseek  `easy-complexsearch` can also be used to search full databases: 
+```
+foldseek databases PDB100 pdb tmp 
+foldseek easy-complexsearch example/1tim.pdb.gz pdb result tmpFolder
+```
+
+#### Complex Search Output
 ##### Tab-separated-complex
-The default fields are containing the following fields: `query,target,fident,alnlen,mismatch,gapopen,qstart,qend,tstart,tend,evalue,bits,complexassignid` but they can be customized with the `--format-output` option e.g. `--format-output "query,target,complexqtmscore,complexttmscore,complexassignid"` returns the query and target accession, the tm scores of complex alignment normalized with query and target lengthes, and assignment id. You can choose many different output columns.
+By default, `easy-complexsearch` outputs the alignment as a tab-separated file. The standard fields include `query, target, fident, alnlen, mismatch, gapopen, qstart, qend, tstart, tend, evalue, bits, complexassignid`. Customize output with the `--format-output` option. For example, `--format-output "query,target,complexqtmscore,complexttmscore,complexassignid"` alters the output to show specific scores and identifiers. 
+
 | Code | Description |
 | --- | --- |
 | **Commons** |
@@ -237,22 +249,31 @@ The default fields are containing the following fields: `query,target,fident,aln
 |complexu       | Rotation matrix of Complex alignment (computed to by TM-score) |
 |complext       | Translation vector of Complex alignment (computed to by TM-score) |
 |complexassignid| Index of Complex alignment |
+
+**Example Output:**
 ```
 1tim.pdb.gz_A   8tim.pdb.gz_A   0.967   247 8   0   1   247 1   247 5.412E-43   1527    0
 1tim.pdb.gz_B   8tim.pdb.gz_B   0.967   247 8   0   1   247 1   247 1.050E-43   1551    0
 ```
-##### Report
-Reports are containing the following fields:
+
+##### Complex Report
+`easy-complexsearch` also generates a report format (prefixed `_report`), which provides a summary ot the inter complex chain matching, including identifiers, chains, TM scores, rotation matrices, translation vectors, and assignment IDs. Reports are containing the following fields:
 | Column | Description |
 | --- | --- |
-| (1,2) | Identifiers for query and target complex |
-| (3,4) | Chains of query complex and target complex |
-| (5,6) | TM scores based on query and target residue length |
+| 1 | Identifiers for query complex |
+| 2 | Identifiers for query complex |
+| 3 | Matched chains of query complex |
+| 4 | Matched chains of target complex |
+| 5 | TM scores normalized by query length |
+| 6 | TM scores normalized by target length |
 | (8,9) | Rotation matrix (u) and Translation vector(t) |
-| (9)   | Assignment id |
+| 9   | Complex Assignment Id |
+
+**Example Output:**
 ```
 1tim.pdb.gz 8tim.pdb.gz A,B A,B 0.98941 0.98941 0.999983,0.000332,0.005813,-0.000373,0.999976,0.006884,-0.005811,-0.006886,0.999959 0.298992,0.060047,0.565875  0
 ```
+
 <!-- 
 ##### Interactive HTML 
 Foldseek can locally generate a search result HTML similiar to the [webserver](https://search.foldseek.com) by specifying the format mode `--format-mode 3`
@@ -301,6 +322,7 @@ foldseek createtsv db db clu clu.tsv
 ### Query centered multiple sequence alignment 
 Foldseek can generate a3m based multiple sequence alignments using the following commands. 
 a3m can be converted to fasta format using [reformat.pl](https://raw.githubusercontent.com/soedinglab/hh-suite/master/scripts/reformat.pl) (`reformat.pl in.a3m out.fas`).
+
 ```
 foldseek createdb example/ targetDB
 foldseek createdb example/ queryDB

diff --git a/lib/3di/CMakeLists.txt b/lib/3di/CMakeLists.txt
@@ -5,6 +5,9 @@ add_library(3di
         structureto3diseqdist.cpp
         )
 mmseqs_setup_derived_target(3di)
-
-target_include_directories(3di PRIVATE ..)  # needed for kerasify/keras_model.h
+target_include_directories(3di
+        PRIVATE 
+        .. # kerasify/keras_model.h
+        ${PROJECT_BINARY_DIR}/generated # encoder_weights_3di.kerasify.h
+)
 add_dependencies(3di local-generated)
diff --git a/lib/mmseqs/src/commons/Parameters.h b/lib/mmseqs/src/commons/Parameters.h
@@ -59,6 +59,8 @@ struct MMseqsParameter {
     }
 };
 
+void initParameterSingleton(void);
+#define DEFAULT_PARAMETER_SINGLETON_INIT void initParameterSingleton() { new Parameters; }
 
 class Parameters {
 public:
@@ -712,13 +714,11 @@ class Parameters {
     static Parameters& getInstance()
     {
         if (instance == NULL) {
-            initInstance();
+            initParameterSingleton();
         }
         return *instance;
     }
-    static void initInstance() {
-        new Parameters;
-    }
+    friend void initParameterSingleton(void);
 
     void setDefaults();
     void initMatrices();

diff --git a/lib/mmseqs/src/mmseqs.cpp b/lib/mmseqs/src/mmseqs.cpp
@@ -1,6 +1,7 @@
 #include "Command.h"
 #include "DownloadDatabase.h"
 #include "Prefiltering.h"
+#include "Parameters.h"
 
 const char* binary_name = "mmseqs";
 const char* tool_name = "MMseqs2";
@@ -19,6 +20,8 @@ void init() {
 }
 void (*initCommands)(void) = init;
 
+DEFAULT_PARAMETER_SINGLETON_INIT
+
 std::vector<DatabaseDownload> externalDownloads = {};
 std::vector<KmerThreshold> externalThreshold = {};
 

diff --git a/lib/mmseqs/src/test/TestAlignment.cpp b/lib/mmseqs/src/test/TestAlignment.cpp
@@ -20,6 +20,7 @@
 #include "Matcher.h"
 
 const char* binary_name = "test_alignment";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 int main (int, const char**) {
     const size_t kmer_size=6;

diff --git a/lib/mmseqs/src/test/TestAlignmentPerformance.cpp b/lib/mmseqs/src/test/TestAlignmentPerformance.cpp
@@ -24,6 +24,7 @@
 #include "StripedSmithWaterman.h"
 
 const char* binary_name = "test_alignmentperformance";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 #define MAX_FILENAME_LIST_FILES 4096
 

diff --git a/lib/mmseqs/src/test/TestAlignmentTraceback.cpp b/lib/mmseqs/src/test/TestAlignmentTraceback.cpp
@@ -18,6 +18,7 @@
 #include "Parameters.h"
 
 const char* binary_name = "test_alignmenttraceback";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 struct scores{
     short H;

diff --git a/lib/mmseqs/src/test/TestAlp.cpp b/lib/mmseqs/src/test/TestAlp.cpp
@@ -43,6 +43,7 @@ Contents: pairwise alignment algorithms
 #include "sls_alignment_evaluer.hpp"
 
 const char* binary_name = "test_alp";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 using namespace Sls;
 using namespace std;

diff --git a/lib/mmseqs/src/test/TestBacktraceTranslator.cpp b/lib/mmseqs/src/test/TestBacktraceTranslator.cpp
@@ -1,5 +1,4 @@
 #include "BacktraceTranslator.h"
-#include "Parameters.h"
 
 const char* binary_name = "test_backtracetranslator";
 

diff --git a/lib/mmseqs/src/test/TestBestAlphabet.cpp b/lib/mmseqs/src/test/TestBestAlphabet.cpp
@@ -14,7 +14,6 @@
 #include "DBReader.h"
 #include "Sequence.h"
 #include "Indexer.h"
-#include "Parameters.h"
 
 const char* binary_name = "test_bestalphabet";
 

diff --git a/lib/mmseqs/src/test/TestCompositionBias.cpp b/lib/mmseqs/src/test/TestCompositionBias.cpp
@@ -9,6 +9,7 @@
 #include "Parameters.h"
 
 const char* binary_name = "test_compositionbias";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 void calcLocalAaBiasCorrection(Sequence* seq, SubstitutionMatrix * m){
     const int windowSize = 40;

diff --git a/lib/mmseqs/src/test/TestDiagonalScoring.cpp b/lib/mmseqs/src/test/TestDiagonalScoring.cpp
@@ -14,6 +14,7 @@
 #include "Parameters.h"
 
 const char* binary_name = "test_diagonalscoring";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 int main (int, const char**) {
     size_t kmer_size = 6;

diff --git a/lib/mmseqs/src/test/TestDiagonalScoringPerformance.cpp b/lib/mmseqs/src/test/TestDiagonalScoringPerformance.cpp
@@ -20,6 +20,7 @@ KSEQ_INIT(int, read)
 #include "Parameters.h"
 
 const char* binary_name = "test_diagonalscoringperformance";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 int main (int, const char**) {
     size_t kmer_size = 6;

diff --git a/lib/mmseqs/src/test/TestKmerGenerator.cpp b/lib/mmseqs/src/test/TestKmerGenerator.cpp
@@ -15,6 +15,7 @@
 #include "Parameters.h"
 
 const char* binary_name = "test_kmergenerator";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 int main (int, const char**) {
     const size_t kmer_size=6;

diff --git a/lib/mmseqs/src/test/TestKmerNucl.cpp b/lib/mmseqs/src/test/TestKmerNucl.cpp
@@ -6,6 +6,7 @@
 #include "Orf.h"
 
 const char* binary_name = "test_kmernucl";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 std::string kmerToSting(size_t idx, int size) {
     char output[32];

diff --git a/lib/mmseqs/src/test/TestKmerScore.cpp b/lib/mmseqs/src/test/TestKmerScore.cpp
@@ -8,6 +8,7 @@
 #include "Parameters.h"
 
 const char* binary_name = "test_kmerscore";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 int main (int, const char**) {
     const size_t kmer_size = 6;

diff --git a/lib/mmseqs/src/test/TestKwayMerge.cpp b/lib/mmseqs/src/test/TestKwayMerge.cpp
@@ -8,7 +8,6 @@
 
 #include "SubstitutionMatrix.h"
 #include "Sequence.h"
-#include "Parameters.h"
 
 const char* binary_name = "test_kwaymerge";
 struct KmerEntry{

diff --git a/lib/mmseqs/src/test/TestMultipleAlignment.cpp b/lib/mmseqs/src/test/TestMultipleAlignment.cpp
@@ -14,6 +14,7 @@
 #include "Parameters.h"
 
 const char* binary_name = "test_multiplealignment";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 int main(int, const char**) {
     Parameters& par = Parameters::getInstance();

diff --git a/lib/mmseqs/src/test/TestPSSM.cpp b/lib/mmseqs/src/test/TestPSSM.cpp
@@ -15,6 +15,7 @@
 #include "MultipleAlignment.h"
 
 const char* binary_name = "test_pssm";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 int main (int, const char**) {
     Parameters& par = Parameters::getInstance();

diff --git a/lib/mmseqs/src/test/TestPSSMPrune.cpp b/lib/mmseqs/src/test/TestPSSMPrune.cpp
@@ -16,6 +16,7 @@
 #include <string.h>
 
 const char* binary_name = "test_pssmprune";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 int main (int, const char**) {
     Parameters& par = Parameters::getInstance();

diff --git a/lib/mmseqs/src/test/TestProfileAlignment.cpp b/lib/mmseqs/src/test/TestProfileAlignment.cpp
@@ -19,6 +19,7 @@
 #include "Parameters.h"
 
 const char* binary_name = "test_profilealignment";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 int main (int, const char**) {
     const size_t kmer_size=6;

diff --git a/lib/mmseqs/src/test/TestReduceMatrix.cpp b/lib/mmseqs/src/test/TestReduceMatrix.cpp
@@ -12,6 +12,7 @@
 #include "Parameters.h"
 
 const char* binary_name = "test_reducematrix";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 int main (int, const char**) {
     const int reductionAlphabetSize = 17;

diff --git a/lib/mmseqs/src/test/TestScoreMatrixSerialization.cpp b/lib/mmseqs/src/test/TestScoreMatrixSerialization.cpp
@@ -5,6 +5,7 @@
 #include "Debug.h"
 
 const char* binary_name = "test_scorematrixserialization";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 int main (int, const char**) {
     Parameters& par = Parameters::getInstance();

diff --git a/lib/mmseqs/src/test/TestSequenceIndex.cpp b/lib/mmseqs/src/test/TestSequenceIndex.cpp
@@ -11,6 +11,7 @@
 #include "Parameters.h"
 
 const char* binary_name = "test_sequenceindex";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 int main (int, const char**) {
     size_t kmer_size = 6;

diff --git a/lib/mmseqs/src/test/TestTanTan.cpp b/lib/mmseqs/src/test/TestTanTan.cpp
@@ -11,6 +11,7 @@
 #include "Parameters.h"
 
 const char* binary_name = "test_tantan";
+DEFAULT_PARAMETER_SINGLETON_INIT
 
 int main (int, const char**) {
     const size_t kmer_size = 6;

diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
@@ -3,7 +3,6 @@ set(HAVE_GCS 0 CACHE BOOL "Have Google Cloud Storage SDK")
 include_directories(commons)
 add_subdirectory(strucclustutils)
 add_subdirectory(commons)
-add_subdirectory(version)
 add_subdirectory(workflow)
 
 add_library(foldseek-framework
@@ -14,7 +13,7 @@ add_library(foldseek-framework
         FoldseekBase.cpp
 )
 mmseqs_setup_derived_target(foldseek-framework)
-target_link_libraries(foldseek-framework version gemmiwrapper 3di pulchra kerasify tmalign block-aligner-c)
+target_link_libraries(foldseek-framework gemmiwrapper 3di pulchra kerasify tmalign block-aligner-c)
 add_dependencies(foldseek-framework local-generated)
 
 if(HAVE_GCS)
@@ -23,7 +22,10 @@ if(HAVE_GCS)
 endif()
 
 if (NOT FOLDSEEK_FRAMEWORK_ONLY)
+        add_subdirectory(version)
         add_executable(foldseek foldseek.cpp)
         mmseqs_setup_derived_target(foldseek foldseek-framework)
+        target_link_libraries(mmseqs-framework version)
+        target_link_libraries(foldseek version)
         install(TARGETS foldseek DESTINATION bin)
 endif()
diff --git a/src/commons/LocalParameters.h b/src/commons/LocalParameters.h
@@ -15,13 +15,10 @@ struct FoldSeekDbValidator : public DbValidator {
 
 class LocalParameters : public Parameters {
 public:
-    static void initInstance() {
-        new LocalParameters;
-    }
     LocalParameters();
     static LocalParameters& getLocalInstance() {
         if (instance == NULL) {
-            initInstance();
+            initParameterSingleton();
         }
         return static_cast<LocalParameters&>(LocalParameters::getInstance());
     }
@@ -137,9 +134,7 @@ class LocalParameters : public Parameters {
 
 
 private:
-
     LocalParameters(LocalParameters const&);
-    ~LocalParameters() {};
     void operator=(LocalParameters const&);
 };
 #endif
diff --git a/src/foldseek.cpp b/src/foldseek.cpp
@@ -1,5 +1,6 @@
 #include <cstddef>
 #include "Command.h"
+#include "LocalParameters.h"
 
 const char* binary_name = "foldseek";
 const char* tool_name = "foldseek";
@@ -17,4 +18,5 @@ void init() {
     registerCommands(&baseCommands);
     registerCommands(&foldseekCommands);
 }
-void (*initCommands)(void) = init;
+void (*initCommands)(void) = init;
+void initParameterSingleton() { new LocalParameters; }