Background

The Paracel GeneMatcher2 system has three main hardware layers. These are the GeneMatcher2 hardware accelerator, a user host machine, and post-processing servers.

The heart of the GeneMatcher2 system is a hardware accelerator containing a pipeline of 9216 parallel processing cells implemented in ASIC (Application Specific Integrated Circuit) technology. It accelerates a number of dynamic programming algorithms, such as Smith-Waterman and HMMer. It contains 430 GB of local disk storage for sequence databases, and a Sun Sparc server running proprietary pipeline and file system management software.

Command-line access to the GeneMatcher2 is available from all of the login hosts. Post-processing is performed on the Paracel BlastMachine cluster. The post-processors perform the alignment step after searches. Formatting of output is done on the user front-end (login) machine.

btk (BioView Toolkit)

btk is the command-line interface to the GM2. The btk program can be used to perform GeneMatcher2 searches and use other utilities. The following usage information is derived from the btk version 4.1 help output:

btk<category>|<command> [arguments...]

<command> is one of the following:

search - Run a search -- "btk search -help" lists searches
qstat - Display queue status

<category> is one of the following:

db - database commands
dbset - database set commands
host - host commands
matrix - matrix commands
param - parameter commands
ppset - post processing set commands
query - query commands
queryset - query set commands
searchset - search set commands
user - user commands

For additional help on a category of commands, type "btk<category> -help".
For additional help on a command, type "btk<command> -help"

btk search commands

The following searchs are accelerated on the GeneMatcher2 hardware using the command btk search <searchname> [<arg1> [<arg2> ...]]:

SMITH-WATERMAN SEARCHES:
swp - protein queries vs protein databases
tswn - protein queries vs DNA (translated .rframe) databases
swx - DNA queries (translated) vs protein databases
swn - DNA queries vs DNA databases
tswx - DNA queries (translated) vs DNA (translated .rframe) databases

HMM AND PROSITE GENERALIZED PROFILE (GP) SEARCHES:
hmm - HMM/GP queries vs protein databases
hframe - HMM/GP queries vs EST/cDNA (translated .rframe) databases
genewise - HMM/GP queries vs genomic DNA (as .codon) databases, allowing for the possibility of introns

GCG PROFILE SEARCHES:
profile - profile queries vs protein databases
pframe - profile queries vs DNA (translated .rframe) databases

The following searches are supported, but are accelerated on the Pentium
cluster rather than the GeneMatcher hardware:

PROSITE REGULAR EXPRESSION SEARCHES:
regexp - Prosite regular expression queries vs protein databases

To invoke any of these searches, type:

<searchname> [<arg1> [<arg2> ...]]

For help on a specific type of search, type "<searchname> -help".

fdf (Fast Data Finder)

fdf is the command-line interface to the GM2's proprietary file system. This file system is UNIX-like, and resides on the actual GM2 hardware. All databases loaded onto the GM2 reside on this filesystem.

FAST DATA FINDER (FDF) FILE MAINTENANCE COMMANDS:

The following commands are invoked by typing only the command name at the command line. You do not need to type btk first.

fdf help - get additional Fast Data Finder help
fdf ls - list a GeneMatcher directory
fdf ls -l - list a GeneMatcher directory, including the number
of sequences and bytes in database)
fdf df - show GeneMatcher disk usage
fdf fdfstat - show status and configuration of GeneMatcher hardware
fdf mkdir - make a directory
fdf rmdir - remove a directory
fdf read - read single sequence from a GeneMatcher database
fdf rm - remove a GeneMatcher database
fdf mv - move or rename a GeneMatcher database