RecBlast is a module that provides a Reciprocal searching framework for finding homologues of sequences between a query species, and a list of target species. RecBlast is designed to be flexible by allowing the user to customize each step of the pipeline, from the forward and reverse search algorithms, to the sequence database for hits. The output of the search is also highly customizable, and can be set to either provide 1:1 Best-Hits or 1:Many hits. The package is written in a modular, parallelized fashion to maximize search efficiency while also allowing for easy additions and modifications to the pipeline.
- Reciprocal searching of any number of sequences between a query species and any number of target species
- Runs on multiple cores
- Uses either BLAT or BLAST for the searches, with room for other options such as HMMER
- Local or remote search host options
- Can use FASTA files, BLAT or BLAST databases, SQL databases, or Ensembl for sequence retreval
The core of the module is the RecSearch object, which organizes the various settings for the reciprocal search. After initializing it with the target and query species, the search types, and the sequence source, one can then change the criteria for the searches or change advanced settings.
from RecBlast.RecBlast import RecSearch
# Initialize the RecSearch Object
# We are using protein sequences from Humans to query the African Elephant genome,
# using BLAT as the algorithm, and obtaining results from the twobit file.
recblast = RecSearch(target_species="Loxodonta africana", query_species="Homo sapiens",
forward_search_type="tblat", reverse_search_type="blat-transcript",
sequence_source="twobit", verbose=0)
# Read query sequences from a FASTA file
recblast.set_queries("Test/query/3protein.fasta", infile_type="fasta")
# For the forward BLAT search, set 'database' to the loxAfr 2bit file
# RecSearch will by default automatically search for a 2bit file otherwise.
recblast.forward_search_settings['database'] = {"Loxodonta africana": "loxAfr3.2bit"}
# For the forward and reverse BLAT searches, set the ports of the gfServer for each species
# By default, it looks for the server at localhost, but it can also be remote if one wishes
recblast.forward_search_settings['database_port'] = {"Loxodonta africana":20008}
recblast.reverse_search_settings['database_port'] = {"Homo sapiens":20005}
# Obtain the sequence for the forward hit from the twobit file.
recblast.sequence_source_settings['database'] = {"Loxodonta africana": "loxAfr3.2bit"}The search is then initialized by running the object, and will return a RecBlastContainer object with all the results and intermediary outputs.
rc = recblast("Example", output_type="BED)The script begins by initializing various RecSearch worker processes, one for each
species-sequence pair, up to a mutable maximum limit (by default, it is half the number of
CPUs detected by multiprocess). Once initialized, each RecSearch process executes the
following protocol:
- Forward Search of the given sequence against the target species database;
- Filter hits based on given parameters;
- Get the full-length sequences of all filtered hits;
- Reverse Search each filtered hit against the query species database;
- Filter reciprocal hits based on given parameters;
- Annotate hits using the filtered, sorted reciprocal hits.
The module has its own function, simply called blast(), that handles the search functionality and calls to BLAST. In addition to being able to specify whether to use a local BLAST installation or NCBI's servers, both the Forward and Reverse BLAST calls take a dictionary parameter, fw_blast_kwargs and rv_blast_kwargs, respectively, that can be used to modify the two searches independently.
- Entrez implementation for remote sequence searches
- Additional tools for analyzing output
RecBlast is a module that provides a Reciprocal searching framework for finding homologues of sequences between a query species, and a list of target species. RecBlast is designed to be flexible by allowing the user to customize each step of the pipeline, from the forward and reverse search algorithms, to the sequence database for hits. The output of the search is also highly customizable, and can be set to either provide 1:1 Best-Hits or 1:Many hits. The package is written in a modular, parallelized fashion to maximize search efficiency while also allowing for easy additions and modifications to the pipeline.
- Reciprocal searching of any number of sequences between a query species and any number of target species
- Runs on multiple cores
- Uses either BLAT or BLAST for the searches, with room for other options such as HMMER
- Local or remote search host options
- Can use FASTA files, BLAT or BLAST databases, SQL databases, or Ensembl for sequence retreval
The core of the module is the RecSearch object, which organizes the various settings for the reciprocal search. After initializing it with the target and query species, the search types, and the sequence source, one can then change the criteria for the searches or change advanced settings.
from RecBlast.RecBlast import RecSearch
# Initialize the RecSearch Object
# We are using protein sequences from Humans to query the African Elephant genome,
# using BLAT as the algorithm, and obtaining results from the twobit file.
recblast = RecSearch(target_species="Loxodonta africana", query_species="Homo sapiens",
forward_search_type="tblat", reverse_search_type="blat-transcript",
sequence_source="twobit", verbose=0)
# Read query sequences from a FASTA file
recblast.set_queries("Test/query/3protein.fasta", infile_type="fasta")
# For the forward BLAT search, set 'database' to the loxAfr 2bit file
# RecSearch will by default automatically search for a 2bit file otherwise.
recblast.forward_search_settings['database'] = {"Loxodonta africana": "loxAfr3.2bit"}
# For the forward and reverse BLAT searches, set the ports of the gfServer for each species
# By default, it looks for the server at localhost, but it can also be remote if one wishes
recblast.forward_search_settings['database_port'] = {"Loxodonta africana":20008}
recblast.reverse_search_settings['database_port'] = {"Homo sapiens":20005}
# Obtain the sequence for the forward hit from the twobit file.
recblast.sequence_source_settings['database'] = {"Loxodonta africana": "loxAfr3.2bit"}The search is then initialized by running the object, and will return a RecBlastContainer object with all the results and intermediary outputs.
rc = recblast("Example", output_type="BED)The script begins by initializing various RecSearch worker processes, one for each
species-sequence pair, up to a mutable maximum limit (by default, it is half the number of
CPUs detected by multiprocess). Once initialized, each RecSearch process executes the
following protocol:
- Forward Search of the given sequence against the target species database;
- Filter hits based on given parameters;
- Get the full-length sequences of all filtered hits;
- Reverse Search each filtered hit against the query species database;
- Filter reciprocal hits based on given parameters;
- Annotate hits using the filtered, sorted reciprocal hits.
The module has its own function, simply called blast(), that handles the search functionality and calls to BLAST. In addition to being able to specify whether to use a local BLAST installation or NCBI's servers, both the Forward and Reverse BLAST calls take a dictionary parameter, fw_blast_kwargs and rv_blast_kwargs, respectively, that can be used to modify the two searches independently.
- Entrez implementation for remote sequence searches
- Additional tools for analyzing output