dbsnp-database

name	dbsnp-database
description	Query NCBI dbSNP for SNP records by rsID, gene, or region via E-utilities and Variation Services REST API. Retrieve alleles, MAF, variant class (SNV/indel/MNV), clinical links, cross-DB IDs (ClinVar, dbVar, 1000G). Free; 3 req/sec (10 with key). For clinical pathogenicity use clinvar-database; for population frequencies use gnomad-database.
license	CC0-1.0

dbSNP Database

Overview

NCBI dbSNP is the primary public repository for short human genetic variants, cataloguing over 1 billion SNPs, indels, and MNVs with allele frequencies, functional annotations, and cross-references to ClinVar, gnomAD, and 1000 Genomes. Variants are identified by stable rsIDs (reference SNP cluster IDs). Access is free via two APIs: the legacy NCBI E-utilities and the newer NCBI Variation Services REST API, which returns structured JSON.

When to Use

• Looking up allele frequencies and variant class for a known rsID
• Searching all dbSNP variants in a gene or chromosomal region by name or coordinates
• Resolving rsIDs to genomic coordinates (GRCh38/GRCh37) and HGVS notation
• Checking whether a variant of interest has clinical significance links to ClinVar entries
• Batch-fetching hundreds of rsIDs efficiently using epost+efetch history server
• Cross-referencing a list of variant positions to dbSNP rsIDs for downstream annotation
• For clinical pathogenicity classifications use clinvar-database; dbSNP provides IDs and frequency but not curated clinical significance
• For population frequency stratified by ancestry use gnomad-database; dbSNP MAF is a single aggregate frequency

Prerequisites

• Python packages: requests, pandas, matplotlib, xml.etree.ElementTree (stdlib)
• Data requirements: rsIDs (rs80357906), gene symbols, or chromosomal coordinates
• Environment: internet connection; NCBI Entrez email required for E-utilities (set email parameter)
• Rate limits: 3 requests/second without API key; 10 requests/second with free NCBI API key. Register at https://www.ncbi.nlm.nih.gov/account/ — add &api_key=YOUR_KEY to all requests

pip install requests pandas matplotlib
# xml.etree.ElementTree is part of Python stdlib — no additional install needed

Quick Start

import requests
import json

EMAIL = "[e-mail]"      # required by NCBI policy
BASE_EUTILS = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils"
BASE_VARIATION = "https://api.ncbi.nlm.nih.gov/variation/v0"

def fetch_snp_by_rsid(rsid: str) -> dict:
    """Fetch a dbSNP record by rsID using the NCBI Variation Services API (structured JSON)."""
    rs_num = str(rsid).lstrip("rs")
    r = requests.get(f"{BASE_VARIATION}/refsnp/{rs_num}", timeout=15)
    r.raise_for_status()
    return r.json()

record = fetch_snp_by_rsid("rs1800497")  # DRD2 Taq1A
print(f"rsID: rs{record['refsnp_id']}")
print(f"Variant type: {record['primary_snapshot_data'].get('variant_type')}")
# Top-level keys: citations, create_date, dbsnp1_merges, last_update_build_id,
# last_update_date, lost_obs_movements, mane_select_ids, present_obs_movements,
# primary_snapshot_data, refsnp_id. (No top-level `organism` field.)
# rsID: rs1800497
# Variant type: snv

Core API

Query 1: rsID Lookup via E-utilities

Fetch the full SNP record for a single rsID using efetch with db=snp. Returns an XML document with alleles, placements, and frequency data.

import requests
import xml.etree.ElementTree as ET

EMAIL = "[e-mail]"
BASE = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils"

def efetch_snp_xml(rsid: str) -> ET.Element:
    """Fetch dbSNP XML record for a single rsID via the docsum rettype.
    Note: rettype="xml" returns a namespaced ExchangeSet; rettype="docsum"
    returns the simpler eSummaryResult/DocumentSummary tree without namespaces."""
    rs_num = str(rsid).lstrip("rs")
    r = requests.get(f"{BASE}/efetch.fcgi",
                     params={"db": "snp", "id": rs_num,
                             "rettype": "docsum", "retmode": "xml",
                             "email": EMAIL},
                     timeout=20)
    r.raise_for_status()
    return ET.fromstring(r.text)

root = efetch_snp_xml("rs80357906")

# Parse the DocumentSummary record (MAF/MAFALLELE were removed in 2024;
# GLOBAL_MAFS is a sub-tree — use ESummary JSON below for easier access)
for docsum in root.iter("DocumentSummary"):
    rs_id = docsum.get("uid")
    snp_class = docsum.findtext("SNP_CLASS", "Unknown")
    chr_pos = docsum.findtext("CHRPOS", "N/A")
    clin_sig = docsum.findtext("CLINICAL_SIGNIFICANCE", "N/A")
    print(f"rs{rs_id} | Class: {snp_class} | Position: ")
    ()

# Fetch using ESummary for structured JSON (preferred for batch)
def esummary_snp(rsid: str) -> dict:
    rs_num = str(rsid).lstrip("rs")
    r = requests.get(f"{BASE}/esummary.fcgi",
                     params={"db": "snp", "id": rs_num,
                             "retmode": "json", "email": EMAIL},
                     timeout=15)
    r.raise_for_status()
    result = r.json()["result"]
    return result.get(rs_num, {})

rec = esummary_snp("rs80357906")
print(f"rs{rec.get('snp_id')}:")
print(f"  Class       : {rec.get('snp_class')}")            # e.g., 'delins'
# `maf`/`mafallele` were removed from ESummary in 2024 — use `global_mafs`
# (list of {study, freq}) and pick a study (e.g., 'GnomAD_genomes') or the
# global aggregate ('TOPMED'/'1000Genomes').
for m in rec.get('global_mafs', [])[:4]:
    print(f"  MAF[{m['study']:18s}]: {m['freq']}")
print(f"  ChrPos      : {rec.get('chrpos')}")               # 17:43057062
print()
()

Query 2: Gene Variant Search

Search dbSNP for all variants in a gene using esearch. Returns a list of rsIDs matching the gene.

import requests

EMAIL = "[e-mail]"
BASE = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils"

def esearch_snp(query: str, retmax: int = 100) -> tuple[list, int]:
    """Search dbSNP using a query string. Returns (id_list, total_count)."""
    r = requests.get(f"{BASE}/esearch.fcgi",
                     params={"db": "snp", "term": query,
                             "retmax": retmax, "retmode": "json",
                             "email": EMAIL},
                     timeout=15)
    r.raise_for_status()
    result = r.json()["esearchresult"]
    return result["idlist"], int(result["count"])

# All variants in BRCA1
ids, total = esearch_snp("BRCA1[gene] AND human[orgn]", retmax=20)
print(f"BRCA1 variants in dbSNP: {total:,} total")
print(f"First 5 rsIDs: {['rs' + i for i in ids[:5]]}")

# Only clinical variants (linked to ClinVar)
ids_clin, total_clin = esearch_snp(
    "BRCA1[gene] AND human[orgn] AND clinsig[filter]", retmax=50)
print(f"BRCA1 variants with clinical significance: ")

Query 3: Chromosomal Region Search

Search for all variants in a genomic region using chromosome coordinates.

import requests

EMAIL = "[e-mail]"
BASE = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils"

def search_region(chrom: str, start: int, stop: int,
                  assembly: str = "GRCh38", retmax: int = 200) -> tuple[list, int]:
    """Find all dbSNP variants in a chromosomal region."""
    query = f"{chrom}[CHR] AND {start}:{stop}[CHRPOS37]" if assembly == "GRCh37" else \
            f"{chrom}[CHR] AND {start}:{stop}[CHRPOS]"
    r = requests.get(f"{BASE}/esearch.fcgi",
                     params={"db": "snp", "term": query,
                             "retmax": retmax, "retmode": "json",
                             "email": EMAIL},
                     timeout=20)
    r.raise_for_status()
    result = r.json()["esearchresult"]
    return result["idlist"], int(result["count"])

# PCSK9 exon 4 region (GRCh38)
ids, total = search_region("1", 55039700, 55040200)
print(f"Variants in chr1:55039700-55040200:  total")
()

Query 4: Variant Summary — MAF, Alleles, Clinical Significance

Retrieve structured summary data for variant records using ESummary, extracting MAF, alleles, and database cross-links.

import requests, json

EMAIL = "[e-mail]"
BASE = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils"

def fetch_snp_summaries(rsids: list) -> dict:
    """Fetch ESummary records for a list of rsIDs. Returns dict keyed by rs number."""
    ids_str = ",".join(str(r).lstrip("rs") for r in rsids)
    r = requests.post(f"{BASE}/esummary.fcgi",
                      data={"db": "snp", "id": ids_str,
                            "retmode": "json", "email": EMAIL},
                      timeout=20)
    r.raise_for_status()
    return r.json()["result"]

rsids = ["rs80357906", "rs80357220", "rs28897672", "rs1801133"]
result = fetch_snp_summaries(rsids)

for rs_num in rsids:
    uid = str(rs_num).lstrip("rs")
    rec = result.get(uid, {})
    # `global_mafs` is a list of {study, freq}; pick the first or filter by study
    gmafs = rec.get("global_mafs", [])
    maf_str = gmafs[0]["freq"] if gmafs else "N/A"
    maf_study = gmafs[0]["study"] if gmafs else ""
    print()
    ()
    ()
    ()
    ()
    ()

Query 5: Batch rsID Query with EPost+EFetch

Efficiently upload hundreds of rsIDs to the NCBI history server using EPost, then retrieve them in batches with EFetch.

import requests, time, pandas as pd

EMAIL = "[e-mail]"
BASE = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils"

def epost_ids(id_list: list) -> tuple[str, str]:
    """Upload rsIDs to NCBI history server. Returns (WebEnv, query_key)."""
    ids_str = ",".join(str(i).lstrip("rs") for i in id_list)
    r = requests.post(f"{BASE}/epost.fcgi",
                      data={"db": "snp", "id": ids_str, "email": EMAIL},
                      timeout=30)
    r.raise_for_status()
    import xml.etree.ElementTree as ET
    root = ET.fromstring(r.text)
    webenv = root.findtext("WebEnv")
    query_key = root.findtext("QueryKey")
    return webenv, query_key

def efetch_history(webenv: str, query_key: str,
                   retstart: int = 0, retmax: int = 100) -> dict:
    """Retrieve records from NCBI history server using ESummary."""
    r = requests.get(f"{BASE}/esummary.fcgi",
                     params={"db": "snp", "WebEnv": webenv,
                             "query_key": query_key, : retstart,
                             : retmax, : ,
                             : EMAIL},
                     timeout=)
    r.raise_for_status()
     r.json()[]

rsid_batch = [, , , ,
              , , ]

webenv, query_key = epost_ids(rsid_batch)
()

records = []
 start  (, (rsid_batch), ):
    result = efetch_history(webenv, query_key, retstart=start, retmax=)
     uid  result.get(, []):
        rec = result[uid]
        
        gmafs = rec.get(, [])
        records.append({
            : ,
            : rec.get(),
            : gmafs[][]  gmafs  ,
            : gmafs[][]  gmafs  ,
            : rec.get(),
            : rec.get(),
        })
    time.sleep()

df = pd.DataFrame(records)
()
(df.to_string(index=))

Query 6: NCBI Variation Services API

The newer REST API returns structured JSON with detailed allele placements, frequencies, and variant type annotations. Preferred for programmatic rsID resolution.

import requests
import pandas as pd

BASE_VARIATION = "https://api.ncbi.nlm.nih.gov/variation/v0"

def fetch_refsnp(rsid: str) -> dict:
    """Fetch structured JSON from NCBI Variation Services API."""
    rs_num = str(rsid).lstrip("rs")
    r = requests.get(f"{BASE_VARIATION}/refsnp/{rs_num}", timeout=15)
    r.raise_for_status()
    return r.json()

def parse_allele_frequencies(record: dict) -> list:
    """Extract allele frequencies from a Variation Services record.
    freq[i] keys: study_name, study_version, local_row_id, observation, allele_count, total_count.
    The asserted allele identifier lives in freq.observation.inserted_sequence (NOT at ann.allele)."""
    freqs = []
    snapshot = record.get("primary_snapshot_data", {})
    for ann in snapshot.get("allele_annotations", []):
        for f in ann.get("frequency", []):
            obs = f.get("observation", {})
            freqs.append({
                "allele": obs.get("inserted_sequence"),
                "study": f.get("study_name"),
                "allele_count": f.get("allele_count"),
                "total_count": f.get("total_count"),
                "freq": (f["allele_count"] / f["total_count"]
                          f.get()  ),
            })
     freqs

record = fetch_refsnp()   
()
()  

placements = record[].get(, [])
 placement  placements[:]:
    seq_id = placement.get()
     allele  placement.get(, []):
        spdi = allele.get(, {}).get(, {})
        ()

freqs = parse_allele_frequencies(record)
 freqs:
    df_freq = pd.DataFrame(freqs[:])
    ()
    (df_freq.to_string(index=))

Key Concepts

rsID vs. ss ID (Submitted SNP)

dbSNP uses two IDs: rs IDs (Reference SNP cluster IDs) are stable public identifiers assigned after clustering submitted variants. ss IDs (Submitted SNP IDs) are assigned to individual laboratory submissions before clustering. Use rs IDs for all queries — ss IDs are internal and submission-specific. A single rs ID may cluster multiple ss IDs from different submissions.

MAF vs. Clinical Significance

• MAF (Minor Allele Frequency): The aggregate frequency of the minor allele across all dbSNP submissions. This is a population-level statistic aggregated from 1000 Genomes, gnomAD, TOPMED, and other studies. It does not indicate whether the variant is pathogenic.
• Clinical significance: A link field pointing to ClinVar classifications (Pathogenic, VUS, Benign, etc.). A variant can have a very low MAF (rare) yet be classified as Benign, or a moderate MAF yet be Pathogenic in a specific context (e.g., founder variants). Use clinvar-database for the full pathogenicity record.

Variant Classes in dbSNP

Class	Description	Example
snv	Single nucleotide variant (A>T)	rs80357906
indel	Insertion or deletion	rs786201005
mnv	Multi-nucleotide variant	rs1057519737
ins	Pure insertion	rs113993960
del	Pure deletion	rs66767301
microsatellite	STR (short tandem repeat)	rs5030655

Common Workflows

Workflow 1: Batch rsID Annotation from a Variant List

Goal: Given a list of rsIDs from a variant call pipeline, retrieve MAF, position, and clinical significance for all variants in one run.

import requests, time, pandas as pd

EMAIL = "[e-mail]"
BASE = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils"

def epost_snp(rsids):
    ids = ",".join(str(r).lstrip("rs") for r in rsids)
    r = requests.post(f"{BASE}/epost.fcgi",
                      data={"db": "snp", "id": ids, "email": EMAIL}, timeout=30)
    r.raise_for_status()
    import xml.etree.ElementTree as ET
    root = ET.fromstring(r.text)
    return root.findtext("WebEnv"), root.findtext("QueryKey")

def esummary_history(webenv, query_key, start, retmax=100):
    r = requests.get(f"{BASE}/esummary.fcgi",
                     params={"db": "snp", "WebEnv": webenv,
                             "query_key": query_key, "retstart": start,
                             "retmax": retmax, "retmode": "json", "email": EMAIL},
                     timeout=30)
    r.raise_for_status()
    return r.json()["result"]

# Example: VCF post-processing — annotate a list of called variants
variant_rsids = [
    "rs80357906", "rs80357220", "rs28897672", ,
    , , , ,
]

()
webenv, query_key = epost_snp(variant_rsids)

records = []
 start  (, (variant_rsids), ):
    result = esummary_history(webenv, query_key, start=start, retmax=)
     uid  result.get(, []):
        rec = result[uid]
        gmafs = rec.get(, [])  
        records.append({
            : ,
            : rec.get(),
            : gmafs[][]  gmafs  ,
            : gmafs[][]  gmafs  ,
            : rec.get(),
            : rec.get(, [{}])[].get()  rec.get()  ,
            : rec.get(),
            : rec.get(),
        })
    time.sleep()

df = pd.DataFrame(records)
df.to_csv(, index=)
()
(df[[, , , , ]].to_string(index=))

Workflow 2: Gene Variant Class Distribution Visualization

Goal: Search all dbSNP variants in a gene and plot their variant class distribution.

import requests, time
import pandas as pd
import matplotlib.pyplot as plt

EMAIL = "[e-mail]"
BASE = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils"

def search_gene_variants(gene: str, retmax: int = 500) -> list:
    r = requests.get(f"{BASE}/esearch.fcgi",
                     params={"db": "snp", "term": f"{gene}[gene] AND human[orgn]",
                             "retmax": retmax, "retmode": "json", "email": EMAIL},
                     timeout=20)
    r.raise_for_status()
    return r.json()["esearchresult"]["idlist"]

def fetch_summaries_batch(ids: list, batch_size: int = 100) -> list:
    records = []
    for i in range(0, len(ids), batch_size):
        batch = ids[i:i+batch_size]
        ids_str = ",".join(batch)
        r = requests.post(f"{BASE}/esummary.fcgi",
                          data={"db": "snp", "id": ids_str,
                                "retmode": , : EMAIL},
                          timeout=)
        r.raise_for_status()
        result = r.json()[]
         uid  result.get(, []):
            rec = result[uid]
            records.append({: , : rec.get(, )})
        time.sleep()
     records

gene = 
()
ids = search_gene_variants(gene, retmax=)
()

records = fetch_summaries_batch(ids)
df = pd.DataFrame(records)
class_counts = df[].value_counts()


fig, ax = plt.subplots(figsize=(, ))
colors = [, , , , , ]
bars = ax.bar(class_counts.index, class_counts.values,
              color=colors[:(class_counts)], edgecolor=)
ax.bar_label(bars, padding=, fontsize=)
ax.set_xlabel()
ax.set_ylabel()
ax.set_title()
plt.tight_layout()
plt.savefig(, dpi=, bbox_inches=)
()
(class_counts.to_string())

Key Parameters

Parameter	Function/Endpoint	Default	Range / Options	Effect
db	All E-utilities	required	"snp"	Database selector; must be "snp" for dbSNP queries
id	efetch, esummary, epost	required	rsID number(s) without rs prefix	Variant identifier(s) to fetch
term	esearch	required	dbSNP query string	Search expression with field tags: [gene], [CHR], [CHRPOS], [rs], clinsig[filter]
retmax	esearch	20	1–10000	Maximum records returned per search
retmode	esearch, esummary	"xml"	"json", "xml"	Response format; use "json" for easy parsing
rettype	efetch	"docsum"	"docsum", "xml"	Record type for efetch responses
WebEnv + query_key	esummary, efetch	—	from epost response	History server tokens for batch retrieval
email	All E-utilities	required	valid email string	NCBI policy; used for rate attribution
api_key	All E-utilities	optional	NCBI API key string	Raises rate limit from 3 to 10 req/sec

Best Practices

1. Register for a free NCBI API key: Adds api_key=YOUR_KEY to requests and triples your rate limit (3 → 10 req/sec) with no other changes. Register at https://www.ncbi.nlm.nih.gov/account/.

2. Use epost+esummary for batches of more than 10 rsIDs: Avoid looping individual efetch calls. EPost uploads all IDs in one request to the history server; subsequent ESummary calls retrieve them in configurable batches of up to 500.

3. Prefer NCBI Variation Services API for structured JSON: The /variation/v0/refsnp/{rs_num} endpoint returns a fully structured JSON with SPDI allele representations, placements, and frequency tables. Easier to parse than E-utilities XML for modern applications.

4. Check snp_class before interpreting MAF: Indels and MNVs use different allele counting conventions than SNVs. Treat multi-allelic sites carefully — the reported MAF may refer to one allele among several.

5. Combine dbSNP with ClinVar lookups: dbSNP records the clinical_significance field as a string (e.g., "pathogenic") but does not contain submitter details, review status, or HGVS details. Use clinvar-database for the full pathogenicity record when clinical interpretation is required.

Common Recipes

Recipe: Quick rsID Existence and Class Check

When to use: Check whether a variant is registered in dbSNP before downstream annotation.

import requests

EMAIL = "[e-mail]"

def check_rsid(rsid: str) -> dict:
    """Check if an rsID exists in dbSNP and return basic info."""
    rs_num = str(rsid).lstrip("rs")
    r = requests.get(
        "https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esummary.fcgi",
        params={"db": "snp", "id": rs_num, "retmode": "json", "email": EMAIL},
        timeout=10
    )
    result = r.json().get("result", {})
    rec = result.get(rs_num, {})
    if not rec or "error" in rec:
        return {"rsid": rsid, "found": False}
    gmafs = rec.get("global_mafs", [])
    return {
        "rsid": rsid,
        "found": True,
        "snp_class": rec.get("snp_class"),
        "chrpos": rec.get("chrpos"),
        "maf": gmafs[0]["freq"] if gmafs else None,
        "maf_study": gmafs[0]["study"] if gmafs else None,
        : rec.get(),
    }

 rsid  [, , ]:
    info = check_rsid(rsid)
     info[]:
        ()
    :
        ()

Recipe: Resolve Gene Variants to rsIDs and Coordinates

When to use: Convert a gene name to a list of rsIDs for use in downstream tools (PLINK, ANNOVAR, etc.).

import requests, time, pandas as pd

EMAIL = "[e-mail]"
BASE = "https://eutils.ncbi.nlm.nih.gov/entrez/eutils"

def gene_to_rsids(gene: str, max_variants: int = 200) -> pd.DataFrame:
    """Search dbSNP for a gene and return rsIDs with GRCh38 coordinates."""
    # Step 1: search
    r = requests.get(f"{BASE}/esearch.fcgi",
                     params={"db": "snp", "term": f"{gene}[gene] AND human[orgn]",
                             "retmax": max_variants, "retmode": "json", "email": EMAIL},
                     timeout=15)
    r.raise_for_status()
    ids = r.json()["esearchresult"]["idlist"]
    if not ids:
        return pd.DataFrame()
    time.sleep(0.4)

    # Step 2: fetch summaries
    r2 = requests.post(f"{BASE}/esummary.fcgi",
                       data={"db": "snp", "id": ",".join(ids),
                             "retmode": "json", "email": EMAIL},
                       timeout=30)
    r2.raise_for_status()
    result = r2.json()["result"]
    rows = []
    for uid in result.get("uids", []):
        rec = result[uid]
        gmafs = rec.get(, [])
        rows.append({
            : ,
            : rec.get(),
            : rec.get(),
            : gmafs[][]  gmafs  ,
        })
     pd.DataFrame(rows)

df = gene_to_rsids(, max_variants=)
()
(df.head().to_string(index=))
df.to_csv(, index=)

Troubleshooting

Problem	Cause	Solution
HTTP 429 or connection refused	Rate limit exceeded (3 req/sec)	Add time.sleep(0.35) between requests; register for API key to get 10 req/sec
ESummary returns {"error": "Invalid uid"}	rsID does not exist in dbSNP	Check rsID spelling; verify with NCBI browser; variant may be a novel call not yet in dbSNP
esearch returns 0 results for a gene	Gene symbol mismatch or missing human[orgn] filter	Try adding AND human[orgn]; check NCBI gene symbol at https://www.ncbi.nlm.nih.gov/gene
KeyError: 'maf' / 'mafallele' in ESummary parsing	Fields removed in the 2024 dbSNP ESummary schema	Use rec["global_mafs"] — list of {"study": ..., "freq": "<allele>=<value>/<count>"}; pick a study (GnomAD_genomes, TOPMED, ALFA) and parse the freq string
global_mafs empty	Variant has no aggregated population frequency in dbSNP	Use gnomAD via gnomad-database directly for population frequencies
root.iter("DocumentSummary") returns 0 with rettype="xml"	EFetch XML root is namespaced ({https://www.ncbi.nlm.nih.gov/SNP/docsum}ExchangeSet)	Use rettype="docsum" (no namespace) or pass the full namespaced tag to iter()
Variation Services API returns 404	rsID not found or wrong URL format	Confirm integer rs number (no rs prefix) in /refsnp/{rs_num} endpoint
EPost XML parsing fails	Non-XML response (rate limit HTML error page)	Check response status code first; add retry logic with time.sleep(1)
Batch efetch returns fewer records than posted	Some rsIDs were merged or retired	Cross-check against NCBI merge history; retired rsIDs redirect to current active rs

Related Skills

• clinvar-database — ClinVar pathogenicity classifications for variants identified by rsID (complement to dbSNP)
• gnomad-database — Population allele frequencies by ancestry group (more detailed than dbSNP MAF)
• gwas-database — GWAS Catalog for SNP-trait associations from published GWAS studies
• ensembl-database — Ensembl REST API for variant consequences and gene annotations
• snpeff-variant-annotation — Annotate VCF files with SnpEff and SnpSift, which adds dbSNP rsIDs and functional predictions

References

• NCBI E-utilities for dbSNP — E-utilities field tags and query syntax specific to dbSNP
• NCBI Variation Services API — REST API documentation and interactive Swagger UI
• Sherry et al., Nucleic Acids Research 2001 — dbSNP original description paper
• NCBI E-utilities Reference — Full E-utilities API reference (applies to all NCBI databases)