proteinortho
6.0.14
|
Variables | |
bool | param_verbose = false |
By default no verbose is printed. More... | |
floattype | param_con_threshold = 0.1 |
as a reference: a chain a-b-c-d has 0.25 More... | |
unsigned int | debug_level = 0 |
Debug stderr level. More... | |
floattype | param_sep_purity = 1e-7 |
as a reference: a-b-c will give +/-0.707107 and 2.34857e-08 More... | |
unsigned int | param_max_nodes = 16777216 |
= 2^24. The absolute maximal number of nodes for a connected component. If exceeded then it is skipped entirely. More... | |
floattype | param_min_species = 1 |
Minimum number of species per connected component. More... | |
string | param_rmgraph = "remove.graph" |
Name of the remove graph, containing all edges that are removed by this clustering algorithm. More... | |
bool | param_useWeights = true |
If true, then weights are used for the calculation of the algebraic connectivity. More... | |
unsigned int | param_minOpenmp = 256 |
the minimum size of a for-loop for openmp to activate (openmp has some initialization costs) More... | |
bool | param_useKmereHeuristic = true |
If true then kmere-heuristic is used. For ambigous splits (e.g. all but one entrie of the fiedler vector are below 0) then the fiedler vector is clustered in k groups instead. This will lead to k new groups instead of 2. See the variables kmereHeuristic_* for more informations when the kmere-heuristic will be activated. More... | |
unsigned int | param_maxRam_inKB = 16777216 |
= 16 GB of memory as default. Decreasing this may lead to more power iterations and thus to a increase in runtime as a trade-off More... | |
bool | param_useLapack = true |
If true, then lapack is used accordingly to param_lapack_power_threshold_d. More... | |
unsigned int | critical_min_nodes = 16777216 |
Depricated. More... | |
const unsigned int | min_iter = 16 |
Minimal number of iterations for power iteration. More... | |
unsigned int | param_max_iter = 8192 |
Maximal number of iterations for power iteration. More... | |
floattype | param_epsilon = 1e-8 |
Epsilon for lapack functions dsyevr/ssyevr as well as the power iteration. Set analog to http://people.sc.fsu.edu/~jburkardt/c_src/power_method/power_method_prb.c. More... | |
const unsigned int | kmereHeuristic_minNodes = 1048576 |
= 2^20. The minimum number of nodes of the connected component for the kmere-heuristic. More... | |
const unsigned int | kmereHeuristic_protPerSpecies = 1 |
The mimum number of proteins per species for the kmere-heuristic. More... | |
const unsigned int | kmereHeuristic_minNumberOfGroups = 3 |
The minimum number of species for the kmere-heuristic. More... | |
const unsigned int | maxUseWeightsNumNodes = 1048576 |
= 2^20. The maximum number of nodes for a connected component such that weights are still used for calculations. If the connected component is too large, the unweighted version is used instead to save time. More... | |
floattype | param_lapack_power_threshold_d = -1 |
The minimum graph density for the power iteration method, lapacks (d|s)syevr is used otherwise. If -1 then the linear function is used instead : If d<10^(-5.2)*n, then lapack otherwise power. More... | |
unsigned int | species_counter = 0 |
Number of species (in the input graph) More... | |
unsigned int | protein_counter = 0 |
Number of proteins total (in the input graph) More... | |
vector< string > | species |
Species ID (number) -> Species name. More... | |
vector< protein > | graph |
Graph containing all protein data (see the class protein for more informations) More... | |
floattype | last_stat = 0 |
For buffering progress stats. More... | |
unsigned int | edges = 0 |
Total number of edges. More... | |
vector< shared_ptr< ofstream > > | graph_clean |
File handler to store graph data. More... | |
vector< int > | reorder_table |
Tells how proteins/species must be sorted. More... | |
unsigned long | graph_ram_total_inKB =0 |
The internal size of the input graph in KB (this will be set in parse_file() ) More... | |
unsigned int | num_cpus =1 |
By default only one core is used (change this with -cpus) More... | |
unsigned int critical_min_nodes = 16777216 |
Depricated.
unsigned int debug_level = 0 |
Debug stderr level.
unsigned int edges = 0 |
Total number of edges.
vector<protein> graph |
Graph containing all protein data (see the class protein for more informations)
vector<shared_ptr<ofstream> > graph_clean |
File handler to store graph data.
unsigned long graph_ram_total_inKB =0 |
The internal size of the input graph in KB (this will be set in parse_file() )
const unsigned int kmereHeuristic_minNodes = 1048576 |
= 2^20. The minimum number of nodes of the connected component for the kmere-heuristic.
const unsigned int kmereHeuristic_minNumberOfGroups = 3 |
The minimum number of species for the kmere-heuristic.
const unsigned int kmereHeuristic_protPerSpecies = 1 |
The mimum number of proteins per species for the kmere-heuristic.
floattype last_stat = 0 |
For buffering progress stats.
const unsigned int maxUseWeightsNumNodes = 1048576 |
= 2^20. The maximum number of nodes for a connected component such that weights are still used for calculations. If the connected component is too large, the unweighted version is used instead to save time.
const unsigned int min_iter = 16 |
Minimal number of iterations for power iteration.
unsigned int num_cpus =1 |
By default only one core is used (change this with -cpus)
floattype param_con_threshold = 0.1 |
as a reference: a chain a-b-c-d has 0.25
floattype param_epsilon = 1e-8 |
Epsilon for lapack functions dsyevr/ssyevr as well as the power iteration. Set analog to http://people.sc.fsu.edu/~jburkardt/c_src/power_method/power_method_prb.c.
floattype param_lapack_power_threshold_d = -1 |
The minimum graph density for the power iteration method, lapacks (d|s)syevr is used otherwise. If -1 then the linear function is used instead : If d<10^(-5.2)*n, then lapack otherwise power.
unsigned int param_max_iter = 8192 |
Maximal number of iterations for power iteration.
unsigned int param_max_nodes = 16777216 |
= 2^24. The absolute maximal number of nodes for a connected component. If exceeded then it is skipped entirely.
unsigned int param_maxRam_inKB = 16777216 |
= 16 GB of memory as default. Decreasing this may lead to more power iterations and thus to a increase in runtime as a trade-off
floattype param_min_species = 1 |
Minimum number of species per connected component.
unsigned int param_minOpenmp = 256 |
the minimum size of a for-loop for openmp to activate (openmp has some initialization costs)
string param_rmgraph = "remove.graph" |
Name of the remove graph, containing all edges that are removed by this clustering algorithm.
floattype param_sep_purity = 1e-7 |
as a reference: a-b-c will give +/-0.707107 and 2.34857e-08
bool param_useKmereHeuristic = true |
If true then kmere-heuristic is used. For ambigous splits (e.g. all but one entrie of the fiedler vector are below 0) then the fiedler vector is clustered in k groups instead. This will lead to k new groups instead of 2. See the variables kmereHeuristic_* for more informations when the kmere-heuristic will be activated.
bool param_useLapack = true |
If true, then lapack is used accordingly to param_lapack_power_threshold_d.
bool param_useWeights = true |
If true, then weights are used for the calculation of the algebraic connectivity.
bool param_verbose = false |
By default no verbose is printed.
unsigned int protein_counter = 0 |
Number of proteins total (in the input graph)
vector<int> reorder_table |
Tells how proteins/species must be sorted.
vector<string> species |
Species ID (number) -> Species name.
unsigned int species_counter = 0 |
Number of species (in the input graph)