The code for these tools are unpublished, primarily because they involved too many BMS-internal connections to cleanly extract it for disclosure. This is a stub repo to just host a few (disclosure approved) screenshots.
Our researchers were generally happy to run their own alignments, but didn't have tools at their disposal to do so. Align This presents a simple-to-use web front end that allowed two or more sequences to be provided, either as accessions or raw Fasta format. Multiple alignment algorithms could be chosen, and power users could specify command line flags. Default output was colorized HTML, with several color schemes available that could be dynamically selected after alignment. The default scheme, "Entropy Gaps", is a custom visualization that highlights "high entropy" columns in the alignment, as well as gaps.
Align This also had a special mode where a second set of "reference" sequences could be provided. Each of the "query" sequences were aligned individually, pair-wise, to each of the references. Each query was then "assigned" to the "best" reference, with results typcially being presented in an Excel workbook. This was popular with our virologists, who could combine their experimental sequences with multiple viral reference strains, letting Align This pick the most likely source strain for each of their sequences.
AlignThis functionality is also available as a Perl API.
Our cloning group was industry-leading, but unexpected differences would still be found in our clones. The question was then: Are these differences due to a PCR artifact, or do they represent known polymorphisms? Genomic Sim4 was designed to allow clone sequences to be compared in the context of a reference genome, with polymorphisms from dbSNP and Thousand Genomes superimposed.
The analysis also showed intronic splice junctions extracted from the genome, along with expected base frequency at each of the 5 (10?) flanking positions.
We maintained a large collection of blast databases, comprising internal, public and vendor-supplied sequences. These databases had varying "refresh" schedules, which resulted in disapointment when queries were run on stale data. Friendly Blast was initially put in place to help centralize our DBs, as well as to clearly indicate the age of each database.
The system evolved to allow simple query submission (file upload, accession list, raw fasta), automated algorithm selection and the availability of multiple output formats, including popular "colorized raw" and Excel workbook reports.
The system would also function as a web service, and supported TSV, XML and JSON result formats.
FriendlyBlast was later extended to include support for IgBlast, including another "colorized raw" format and an Excel report. These searches were popular with colleagues designing our antibody-based therapies.
I managed to leave BMS without a screenshot for Friendly Blast ...