The Pfam database is a large collection of protein domain families. Each family is represented by multiple sequence alignments and a hidden Markov model (HMMs).

Here are two solution to setting up protein domain searches:

1) You can set up a local database for use with HMMER, which is the same program Pfam site uses to search for protein domains in submitted queries. You have to download the HMM models, then prepare the database with hmmpress, and use hmmscan or hmmsearch to search the database.

wget ftp://ftp.ebi.ac.uk/pub/databases/Pfam/releases/Pfam31.0/Pfam-A.hmm.gz
gunzip Pfam-A.hmm.gz
hmmpress Pfam-A.hmm
hmmsearch --tblout out.txt -E 1e-5 --cpu 2 Pfam-A.hmm input_proteins.faa

2) RPS-BLAST uses databases compiled from position-specif scoring matrices. The Conserved Domain Database encompasses data from several other databases, Pfam being just one of them - read more here. NCBI provides a pre-formated RPS-BLAST database derived from Pfam:

ftp://ftp.ncbi.nih.gov/pub/mmdb/cdd/little_endian/Pfam_LE.tar.gz

You can use this (and the others available) to set-up a local RPS-BLAST databse.

Not sure if this question is actual but I will answer anyway.

Here is the link to fasta aa Pfam database:

ftp://ftp.ebi.ac.uk/pub/databases/Pfam/current_release/Pfam-A.fasta.gz

You can create local database by running this command (first download blast-ncbi+ (sudo apt-get install blast-ncbi+):

makeblastdb -in Pfam-A.fasta -dbtype prot

Then search your aa sequences against Pfam database. Eg. of blastp below:

blastp -query file.fasta -task blastp-fast -db Pfam-A.fasta -out blast_results.txt -evalue 0.001 -max_target_seqs 1 -num_threads 8 -outfmt '6 qaccver qlen pident length evalue stitle  ' > blast_log.txt

where:

-max_target_seqs 1 - show only one best hit

One line of output:

orf1    93  100.000 93  1.97e-62    A0A087VDV9_BALRE/1-211 A0A087VDV9.1 PF00244.19;14-3-3;

Best, Agata