diff --git a/CMakeLists.txt b/CMakeLists.txt
index c6e1641..71f7e17 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -17,8 +17,8 @@ set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR})
 MESSAGE(STATUS "Compiling for processor: " ${CMAKE_HOST_SYSTEM_PROCESSOR})
 
 if (UNIX AND (CMAKE_HOST_SYSTEM_PROCESSOR STREQUAL "x86_64"))
-  MESSAGE(STATUS "Compiling with flags: -march=native -mbmi2 -msse4.2")
-  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -march=native")
+  MESSAGE(STATUS "Compiling with flags: -march=core-avx2 -mbmi2 -msse4.2")
+  set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -march=core-avx2")
   # Flags for PTHash:
   set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -mbmi2 -msse4.2") # for hardware popcount and pdep
 endif()
diff --git a/include/buckets.hpp b/include/buckets.hpp
index 743c0d0..6954124 100644
--- a/include/buckets.hpp
+++ b/include/buckets.hpp
@@ -5,6 +5,11 @@
 #include "ef_sequence.hpp"
 
 namespace sshash {
+struct superkmer_result {
+    uint64_t kmer_idx;
+    uint64_t superkmer_idx;
+    uint64_t superkmer_id;
+};
 
 template <class kmer_t>
 struct buckets {
@@ -108,6 +113,26 @@ struct buckets {
         }
         return lookup_result();
     }
+    // superkmer annotation
+    lookup_result superkmer_id_to_kmer_id(uint64_t super_kmer_id, uint64_t k) const {
+        uint64_t offset = offsets.access(super_kmer_id);
+        auto [res, contig_end] = offset_to_id(offset, k);
+        return res;
+    }
+    
+
+    superkmer_result lookup_superkmer_start(uint64_t begin, uint64_t end, kmer_t target_kmer,
+                                   uint64_t k, uint64_t m) const {
+        for (uint64_t super_kmer_id = begin; super_kmer_id != end; ++super_kmer_id) {
+            auto res = lookup_in_super_kmer(super_kmer_id, target_kmer, k, m);
+            if(res.kmer_id != constants::invalid_uint64){
+                assert(is_valid(res));
+                return {res.kmer_id, superkmer_id_to_kmer_id(super_kmer_id, k).kmer_id, super_kmer_id};
+            }
+        }
+        return {constants::invalid_uint64, constants::invalid_uint64, constants::invalid_uint64};
+    }
+
 
     lookup_result lookup_canonical(uint64_t bucket_id, kmer_t target_kmer, kmer_t target_kmer_rc,
                                    uint64_t k, uint64_t m) const {
diff --git a/include/builder/util.hpp b/include/builder/util.hpp
index 59cedd3..80201e5 100644
--- a/include/builder/util.hpp
+++ b/include/builder/util.hpp
@@ -219,7 +219,10 @@ struct minimizers_tuples {
     }
 
     std::string get_minimizers_filename() const {
-        assert(m_num_files_to_merge > 0);
+        //assert(m_num_files_to_merge > 0);
+        if(!(m_num_files_to_merge > 0)){
+            throw std::invalid_argument("m_num_files_to_merge > 0");
+        }
         if (m_num_files_to_merge == 1) return get_tmp_output_filename(0);
         std::stringstream filename;
         filename << m_tmp_dirname << "/sshash.tmp.run_" << m_run_identifier << ".minimizers.bin";
diff --git a/include/dictionary.hpp b/include/dictionary.hpp
index 82a68cd..e1f5863 100644
--- a/include/dictionary.hpp
+++ b/include/dictionary.hpp
@@ -1,10 +1,13 @@
 #pragma once
 
+#include <functional>
+
 #include "util.hpp"
 #include "minimizers.hpp"
 #include "buckets.hpp"
 #include "skew_index.hpp"
 #include "weights.hpp"
+#include <mutex>
 
 namespace sshash {
 
@@ -30,6 +33,7 @@ struct dictionary {
     uint64_t k() const { return m_k; }
     uint64_t m() const { return m_m; }
     uint64_t num_contigs() const { return m_buckets.pieces.size() - 1; }
+    uint64_t num_superkmers() const { return m_buckets.offsets.size(); }
     bool canonicalized() const { return m_canonical_parsing; }
     bool weighted() const { return !m_weights.empty(); }
 
@@ -127,6 +131,17 @@ struct dictionary {
     void print_space_breakdown() const;
     void compute_statistics() const;
 
+    // pre: monochromatic_labels returns true iff same labels are found for all k-mers in given string 
+    // post: returns indeces of superkmers that are not monochromatic
+    std::vector<uint64_t> build_superkmer_bv(const std::function<bool (std::string_view)> &monochromatic_labels) const;
+    // pre: get_annotation_labels returns vector of labels for given k-mer
+    // post: number of k-mers and color changes for every super-k-mer are written to files
+    void make_superkmer_stats(const std::function<std::vector<std::string> (std::string_view)> &get_annotation_labels) const;
+
+    superkmer_result kmer_to_superkmer_idx(char const* kmer_str, bool check_reverse_complement) const ;
+    superkmer_result kmer_to_superkmer_idx_helper(kmer_t uint_kmer) const ;
+    uint64_t look_up_from_superkmer_id(uint64_t superkmer_id, char const* kmer_str, bool check_reverse_complement) const ;
+
     template <typename Visitor>
     void visit(Visitor& visitor) const {
         visit_impl(visitor, *this);
diff --git a/include/dictionary.impl b/include/dictionary.impl
index 72d41a9..e3ddabf 100644
--- a/include/dictionary.impl
+++ b/include/dictionary.impl
@@ -1,5 +1,47 @@
 namespace sshash {
 
+template <class kmer_t>
+superkmer_result dictionary<kmer_t>::kmer_to_superkmer_idx_helper(kmer_t uint_kmer) const {
+    uint64_t minimizer = util::compute_minimizer(uint_kmer, m_k, m_m, m_seed);
+    uint64_t bucket_id = m_minimizers.lookup(minimizer);
+    auto [begin, end] = m_buckets.locate_bucket(bucket_id);
+
+    if (m_skew_index.empty()) return m_buckets.lookup_superkmer_start(begin, end, uint_kmer, m_k, m_m);
+
+    uint64_t num_super_kmers_in_bucket = end - begin;
+    uint64_t log2_bucket_size = util::ceil_log2_uint32(num_super_kmers_in_bucket);
+    // superkmer in skew index
+    if (log2_bucket_size > m_skew_index.min_log2) {
+        uint64_t pos = m_skew_index.lookup(uint_kmer, log2_bucket_size);
+        /* It must hold pos < num_super_kmers_in_bucket for the kmer to exist. */
+        if (pos < num_super_kmers_in_bucket) {
+            auto res_kmer = m_buckets.lookup_in_super_kmer(begin + pos, uint_kmer, m_k, m_m).kmer_id;
+            if(res_kmer != sshash::constants::invalid_uint64){
+                uint64_t superkmer_id = begin + pos;
+                return {res_kmer, m_buckets.superkmer_id_to_kmer_id(superkmer_id, m_k).kmer_id, superkmer_id}; 
+            }
+        }
+        return {constants::invalid_uint64, constants::invalid_uint64, constants::invalid_uint64};
+    }
+    // superkmer not in skew index
+    return  m_buckets.lookup_superkmer_start(begin, end, uint_kmer, m_k, m_m);
+
+}
+
+template <class kmer_t>
+uint64_t dictionary<kmer_t>::look_up_from_superkmer_id(uint64_t superkmer_id, char const* kmer_str, bool check_reverse_complement) const {
+    kmer_t uint_kmer = util::string_to_uint_kmer<kmer_t>(kmer_str, m_k);
+
+    auto res = m_buckets.lookup_in_super_kmer(superkmer_id, uint_kmer, m_k, m_m).kmer_id;
+    if (check_reverse_complement and res == constants::invalid_uint64) {
+        kmer_t uint_kmer_rc = uint_kmer;
+        uint_kmer_rc.reverse_complement_inplace(m_k);
+        res = m_buckets.lookup_in_super_kmer(superkmer_id, uint_kmer_rc, m_k, m_m).kmer_id;
+    }
+    return res;
+}
+
+
 template <class kmer_t>
 lookup_result dictionary<kmer_t>::lookup_uint_regular_parsing(kmer_t uint_kmer) const {
     auto minimizer = util::compute_minimizer(uint_kmer, m_k, m_m, m_seed);
@@ -67,6 +109,17 @@ uint64_t dictionary<kmer_t>::lookup_uint(kmer_t uint_kmer, bool check_reverse_co
     return res.kmer_id;
 }
 
+template <class kmer_t>
+superkmer_result dictionary<kmer_t>::kmer_to_superkmer_idx(char const* kmer_str, bool check_reverse_complement) const {
+    kmer_t uint_kmer = util::string_to_uint_kmer<kmer_t>(kmer_str, m_k);
+    superkmer_result sk_res = kmer_to_superkmer_idx_helper(uint_kmer);
+    if(sk_res.kmer_idx == constants::invalid_uint64 && check_reverse_complement){
+        kmer_t uint_kmer_rc = uint_kmer;
+        uint_kmer_rc.reverse_complement_inplace(m_k);
+        sk_res = kmer_to_superkmer_idx_helper(uint_kmer_rc);
+    }
+    return sk_res;
+}
 template <class kmer_t>
 lookup_result dictionary<kmer_t>::lookup_advanced(char const* string_kmer,
                                                   bool check_reverse_complement) const {
diff --git a/include/statistics.impl b/include/statistics.impl
index c9fef8e..e5b7518 100644
--- a/include/statistics.impl
+++ b/include/statistics.impl
@@ -1,5 +1,9 @@
+#include <algorithm>
+#include <functional>
+#include <omp.h>
+#include "dictionary.hpp"
 #include "buckets_statistics.hpp"
-
+#include <time.h> 
 namespace sshash {
 
 template <class kmer_t>
@@ -11,8 +15,7 @@ void dictionary<kmer_t>::compute_statistics() const {
     buckets_statistics buckets_stats(num_minimizers, num_kmers, num_super_kmers);
 
     std::cout << "computing buckets statistics..." << std::endl;
-
-    for (uint64_t bucket_id = 0; bucket_id != num_minimizers; ++bucket_id) {
+    for (uint64_t bucket_id = 0; bucket_id < num_minimizers; ++bucket_id) {
         auto [begin, end] = m_buckets.locate_bucket(bucket_id);
         uint64_t num_super_kmers_in_bucket = end - begin;
         buckets_stats.add_num_super_kmers_in_bucket(num_super_kmers_in_bucket);
@@ -48,5 +51,185 @@ void dictionary<kmer_t>::compute_statistics() const {
     buckets_stats.print_full();
     std::cout << "DONE" << std::endl;
 }
+inline bool equal(const std::vector<std::string>& input1, const std::vector<std::string>& input2) {
+    if(input1.size() != input2.size()){
+        return false;
+    }
+    for(size_t i = 0; i < input1.size(); i++){
+        if(input1[i] != input2[i]){
+            return false;
+        }
+    }
+    return true; 
+}
+
+template <class kmer_t>
+void uint_kmer_to_last_char(kmer_t x, char* str, uint64_t k) {
+    x.drop_chars(k - 1);
+    str[0] = kmer_t::uint64_to_char(x.pop_char());
+}
+
+template <class kmer_t>
+std::vector<uint64_t> dictionary<kmer_t>::build_superkmer_bv(const std::function <bool (std::string_view)> &monochromatic_labels) const {
+    //uint64_t num_kmers = size();
+    uint64_t num_minimizers = m_minimizers.size();
+    //uint64_t num_super_kmers = m_buckets.offsets.size();
+
+    std::cout << "building super kmer mask..." << std::endl;
+    //std::vector<bool> non_mono_superkmer (num_super_kmers, false);
+    //size_t superkmer_idx = 0;
+    uint64_t one_pc_buckets = std::ceil(num_minimizers/100.0);
+    std::cout<<"iterating through buckets :\n";
+    uint64_t num_threads = std::thread::hardware_concurrency();
+    if (num_minimizers < num_threads) num_threads = num_minimizers;
+    std::vector<std::vector<uint64_t>> indeces (num_threads);
+    time_t my_time = time(NULL);
+    printf("%s", ctime(&my_time)); 
+    #pragma omp parallel for num_threads(num_threads) schedule(dynamic)
+    for (uint64_t bucket_id = 0; bucket_id < num_minimizers; ++bucket_id) {
+        if(bucket_id%one_pc_buckets==0){
+	    my_time = time(NULL);
+	    printf("%s", ctime(&my_time));
+	    std::cout <<" "<< bucket_id/one_pc_buckets <<"%" << std::endl;
+	}
+	    auto [begin, end] = m_buckets.locate_bucket(bucket_id);
+        //uint64_t num_super_kmers_in_bucket = end - begin;
+        for (uint64_t super_kmer_id = begin; super_kmer_id != end; ++super_kmer_id) {
+	    uint64_t offset = m_buckets.offsets.access(super_kmer_id);
+            auto [_, contig_end] = m_buckets.offset_to_id(offset, m_k);
+            (void)_;
+            bit_vector_iterator<kmer_t> bv_it(m_buckets.strings, kmer_t::bits_per_char * offset);
+            uint64_t window_size = std::min<uint64_t>(m_k - m_m + 1, contig_end - offset - m_k + 1);
+            kmer_t prev_minimizer = constants::invalid_uint64;
+            std::vector<std::string> prev_labels = {};
+            std::string superkmer = "";
+            size_t num_chars_to_add = m_k;
+            for (uint64_t w = 0; w != window_size; ++w) {
+                kmer_t kmer = bv_it.read_and_advance_by_char(m_k);
+                auto [minimizer, pos] = util::compute_minimizer_pos(kmer, m_k, m_m, m_seed);
+                
+                //check if superkmer has ended
+                if (prev_minimizer != constants::invalid_uint64 and minimizer != prev_minimizer) {
+                    break;
+                }
+                prev_minimizer = minimizer;
+
+                //get kmers into superkmer
+                std::string chars_to_add(num_chars_to_add, '_');
+                if(num_chars_to_add > 1){
+                    util::uint_kmer_to_string(kmer, &chars_to_add[0], m_k);
+                    num_chars_to_add = 1;
+                }else{
+                    uint_kmer_to_last_char(kmer, &chars_to_add[0], m_k);
+                }
+                superkmer += chars_to_add;
+                
+            }   
+            // get labels and compare them to previous ones
+            if(!monochromatic_labels(superkmer)){
+            	indeces[omp_get_thread_num()].push_back(super_kmer_id);
+	    }
+            
+        }
+    }
+    std::cout << "DONE" << std::endl;
+    my_time = time(NULL);
+    printf("%s", ctime(&my_time)); 
+    std::cout << " Merging index vectors...\n";
+    std::sort(indeces.begin(),indeces.end(),[](const std::vector<uint64_t>& first, const std::vector<uint64_t>& second) {return first.size() < second.size(); });
+    std::vector<uint64_t> merged = std::move(indeces[0]);
+    for(uint64_t i = 1; i < num_threads; i++){
+    	std::vector<uint64_t> dest_aux;
+        auto& src = indeces[i];
+        std::merge(merged.begin(),merged.end(),src.begin(), src.end(),std::back_inserter(dest_aux));
+        merged = std::move(dest_aux);
+    }
+    my_time = time(NULL);
+    printf("%s", ctime(&my_time)); 
+    std::cout<< " merged! \n";
+    return merged;
+}
+static void write_vec(const std::vector<unsigned int>& vec, std::string output_path){
+    std::ofstream outputFile(output_path);
+    if (!outputFile.is_open()) {
+        std::cerr << "Error opening the file." << std::endl;
+    }
+    for (const auto& element : vec) {
+        outputFile << element << " ";
+    }
+    outputFile.close();
+}
+template <class kmer_t>
+void dictionary<kmer_t>::make_superkmer_stats(const std::function<std::vector<std::string> (std::string_view)> &get_annotation_labels) const {
+    uint64_t num_minimizers = m_minimizers.size();
+    uint64_t num_super_kmers = m_buckets.offsets.size();
+
+    std::cout << "gathering super-k-mer stats..." << std::endl;
+    std::vector<unsigned> kmer_per_superkmer (num_super_kmers, 0);
+    std::vector<unsigned> color_changes_superkmer (num_super_kmers, 0);
+
+    uint64_t num_threads = std::thread::hardware_concurrency();
+    if (num_minimizers < num_threads) num_threads = num_minimizers;
+    std::mutex color_lock;
+    std::mutex kmer_lock;
+
+    uint64_t one_pc_buckets = std::ceil(num_minimizers/100.0);
+    time_t my_time = time(NULL);
+    std::cout<<"iterating through buckets :\n";
+    #pragma omp parallel for num_threads(num_threads) schedule(dynamic)
+    for (uint64_t bucket_id = 0; bucket_id < num_minimizers; ++bucket_id) {
+        if(bucket_id%one_pc_buckets==0){
+            my_time = time(NULL);
+            printf("%s", ctime(&my_time));
+            std::cout <<" "<< bucket_id/one_pc_buckets <<"%" << std::endl;
+	    }
+	    auto [begin, end] = m_buckets.locate_bucket(bucket_id);
+        for (uint64_t super_kmer_id = begin; super_kmer_id != end; ++super_kmer_id) {
+	        uint64_t offset = m_buckets.offsets.access(super_kmer_id);
+            auto [_, contig_end] = m_buckets.offset_to_id(offset, m_k);
+            (void)_;
+            bit_vector_iterator<kmer_t> bv_it(m_buckets.strings, kmer_t::bits_per_char * offset);
+            uint64_t window_size = std::min<uint64_t>(m_k - m_m + 1, contig_end - offset - m_k + 1);
+            kmer_t prev_minimizer = constants::invalid_uint64;
+            std::vector<std::string> prev_labels = {};
+
+            unsigned num_kmers = 0;
+            unsigned num_col = 0;
+            for (uint64_t w = 0; w != window_size; ++w) {
+                kmer_t kmer = bv_it.read_and_advance_by_char(m_k);
+                auto [minimizer, pos] = util::compute_minimizer_pos(kmer, m_k, m_m, m_seed);
+                
+		        //check if superkmer has ended
+                if (prev_minimizer != constants::invalid_uint64 and minimizer != prev_minimizer) {
+                    break;
+                }
+		        prev_minimizer = minimizer;
+                num_kmers++;
+
+                // get labels and compare them to previous ones
+                std::string kmer_str(m_k,'_');
+                util::uint_kmer_to_string(kmer, &kmer_str[0], m_k);
+                std::vector<std::string> labels = get_annotation_labels(kmer_str);
+            
+                if(prev_labels.size() == 0){
+                    prev_labels = labels;
+                } else if(!equal(labels, prev_labels)){
+                    num_col++;
+                }
+            }
+            color_lock.lock();
+            color_changes_superkmer[super_kmer_id] = num_col;
+            color_lock.unlock();
+            kmer_lock.lock();
+            kmer_per_superkmer[super_kmer_id] = num_kmers;
+            kmer_lock.unlock();
+        }
+    }
+    std::cout << "DONE" << std::endl;
+
+    std::cout<<"writing kmer and color stats to files...\n";
+    write_vec(color_changes_superkmer,"final_color_stats.txt");
+    write_vec(kmer_per_superkmer,"final_kmer_stats.txt");
+}
 
 }  // namespace sshash