All the blast results were extracted without cutoff parameters using Blast Parser (v1.2.6.14; http://geneproject.altervista.org/) and annotated contigs were termed transcripts.
We used the two parameters recently defined by Salse et al. [ 41- 43] to increase the stringency and significance of Blast sequence alignment by parsing Blast results and rebuilding HSPs (High Scoring Pairs) or pairwise sequence alignments to identify accurate paralogous and orthologous relationships.
Results of the BLAST were parsed to identify instances where a gene's BLAST result produced a hit for one and only one of each genome in the analysis.
These stringency criteria for parsing BLAST output are default parameters, but users have the option to specify alternative criteria.
The automation scripts for parsing BLAST [ 14] and HMMER [ 15] output are written in Ruby [ 16] making use of the BioRuby library [ 17].
For example, BioPerl (http://www.bioperl.org) and Biopython (http://biopython.org/) are widely adopted for the tasks of parsing BLAST output and querying the GENBANK database (2, 3).
A semi-automated process of parsing BLAST hits and manually curating the putative annotations resulted in a spreadsheet of information including cloneID, contig number, number of clones in each contig, nearest BLASTX match, accession number of match, length of match, percent similarity, putative annotation and functional classifications.
Each step of the pipeline (aligning probes to transcripts, parsing blast result, identifying transcript specific probe-clusters, creating transcript specific CDF file, processing using R, retrieving and displaying of differentially regulated or detected alternatively spliced forms) was checked manually to ensure that the software works as intended.
First, a pseudo multiple-sequence alignment (MSA) of the query and BLAST hit sequences is created by parsing the standard BLAST output: individual BLAST-aligned hit sequences are piled up by positioning each residue relative to its homologue in the query sequence (high scoring pairs [HSP] stacking).
The candidates are searched with a PSSM and against a local version of KinBase using BLAST, and the BLAST results are parsed.
