Remote BLAST guide

Run BLAST against NCBI's hosted databases with the -remote flag instead of shipping and indexing data yourself. Query size and rate matter most — NCBI throttles remote requests, so keep queries small and add sleeps between submissions to avoid HTTP 429s.

When to use remote BLAST

• Quick public searches without shipping databases.
• Low-frequency runs where download/build overhead dominates.
• Not ideal for large/private databases (prefer local).

Minimal remote command

blastn \
  -query inputs/query.fasta \
  -db nt \
  -remote \
  -entrez_query "Homo sapiens[Organism]" \
  -out outputs/remote.tsv \
  -outfmt "6 qseqid sseqid pident length evalue bitscore"

Throttling and retries

for q in inputs/queries/*.fasta; do
  blastn -query "$q" -db nt -remote \
    -out "outputs/$(basename "$q").tsv" \
    -outfmt "6 std qcovs" \
    || { echo "retrying $q" && sleep 5 && blastn -query "$q" -db nt -remote -out "outputs/$(basename "$q").tsv" -outfmt "6 std qcovs"; }
  sleep 1
done

• Add sleeps between requests to avoid HTTP 429.
• Capture stderr to logs/remote.log for diagnosis.

Tips

• Keep queries small; split large FASTA into many small files.
• Record the Entrez query, database name, and BLAST version in SUMMARY.json.
• If remote latency is high or throttling frequent, switch to local databases.