Publication Intelligence

Scientific publications reveal competitive research directions years before patents or clinical trials appear. By connecting publications with authors, institutions, genes, diseases, and clinical variants from databases like ClinVar, companies can identify emerging therapeutic targets, track institutional expertise, and discover collaboration opportunities. This intelligence becomes particularly valuable in pharma and biotech, where publications often signal strategic intent years before clinical data or product launches, allowing companies to anticipate competitive dynamics in specific therapeutic areas and adjust their development priorities accordingly.

Scenario

A pharma Competitive Intelligence team monitoring Alzheimer’s research faces information overload—thousands of disconnected PubMed results make it impossible to answer strategic questions like: "Which institutions are leading TREM2 research?" or "What new gene-disease links are validated by clinical evidence?" They need to connect these publications to patents, clinical trials, and other research to understand the competitive landscape and identify potential collaboration opportunities.

Solution

The graph connects publications, authors, institutions, genes, proteins, diseases and SNPs in a unified model. This enables traversing complex paths like "Institution, Author, Publication, Gene, Clinical Variant, Disease" in single queries, revealing patterns invisible in traditional databases.

Data Model

Publication intelligence data model

Demo Data

This demo dataset includes:

  • 6 academic and industry institutions

  • 6 authors (with h-index and expertise)

  • 6 top journals

  • major neurodegenerative diseases

  • 6 genes and proteins

  • 6 SNPs (clinical variants from ClinVar)

  • 6 keywords (research themes)

  • 6 publications

  • 6 citations

  • 6 clinical variants

  • 6 diseases

  • 6 genes

  • 6 proteins

  • 6 SNPs

  • 6 keywords

// ============================================
// MERGE INSTITUTIONS (Academic & Pharma)
// ============================================

MERGE (mit:Institution {name: 'MIT', type: 'Academic', country: 'US'})
MERGE (stanford:Institution {name: 'Stanford University', type: 'Academic', country: 'US'})
MERGE (ucl:Institution {name: 'University College London', type: 'Academic', country: 'UK'})
MERGE (dzne:Institution {name: 'DZNE', type: 'Academic', country: 'DE'})
MERGE (genentechRes:Institution {name: 'Genentech Research', type: 'Industry', country: 'US'})
MERGE (biogenRes:Institution {name: 'Biogen Research', type: 'Industry', country: 'US'})

// ============================================
// MERGE AUTHORS
// ============================================

MERGE (chen:Author {name: 'Dr. Li Chen', h_index: 45, expertise: 'Neurogenetics'})
MERGE (rodriguez:Author {name: 'Dr. Maria Rodriguez', h_index: 52, expertise: 'Immunology'})
MERGE (tanaka:Author {name: 'Dr. Kenji Tanaka', h_index: 38, expertise: 'Structural Biology'})
MERGE (schmidt:Author {name: 'Dr. Anna Schmidt', h_index: 41, expertise: 'Neuroscience'})
MERGE (williams:Author {name: 'Dr. James Williams', h_index: 35, expertise: 'Genomics'})
MERGE (kumar:Author {name: 'Dr. Priya Kumar', h_index: 29, expertise: 'Clinical Genetics'})

// Author-Institution relationships
MERGE (chen)-[:AFFILIATED_WITH]->(mit)
MERGE (rodriguez)-[:AFFILIATED_WITH]->(stanford)
MERGE (tanaka)-[:AFFILIATED_WITH]->(genentechRes)
MERGE (schmidt)-[:AFFILIATED_WITH]->(dzne)
MERGE (williams)-[:AFFILIATED_WITH]->(ucl)
MERGE (kumar)-[:AFFILIATED_WITH]->(biogenRes)

// ============================================
// MERGE JOURNALS
// ============================================

MERGE (nature:Journal {name: 'Nature', impact_factor: 49.9})
MERGE (cell:Journal {name: 'Cell', impact_factor: 41.6})
MERGE (natGenet:Journal {name: 'Nature Genetics', impact_factor: 31.7})
MERGE (neuron:Journal {name: 'Neuron', impact_factor: 16.2})

// ============================================
// MERGE DISEASES
// ============================================

MERGE (ad:Disease {name: "Alzheimer's Disease", icd10: 'G30', prevalence: '6.7M US'})
MERGE (pd:Disease {name: "Parkinson's Disease", icd10: 'G20', prevalence: '1M US'})
MERGE (ftd:Disease {name: 'Frontotemporal Dementia', icd10: 'G31.09', prevalence: '60K US'})

// ============================================
// MERGE GENES AND PROTEINS
// ============================================

MERGE (trem2_gene:Gene {symbol: 'TREM2', name: 'Triggering receptor expressed on myeloid cells 2', chromosome: '6'})
MERGE (apoe_gene:Gene {symbol: 'APOE', name: 'Apolipoprotein E', chromosome: '19'})
MERGE (app_gene:Gene {symbol: 'APP', name: 'Amyloid precursor protein', chromosome: '21'})
MERGE (mapt_gene:Gene {symbol: 'MAPT', name: 'Microtubule associated protein tau', chromosome: '17'})
MERGE (bin1_gene:Gene {symbol: 'BIN1', name: 'Bridging integrator 1', chromosome: '2'})

MERGE (trem2_prot:Protein {name: 'TREM2', uniprot: 'Q9NZC2', function: 'Immune receptor'})
MERGE (apoe_prot:Protein {name: 'ApoE', uniprot: 'P02649', function: 'Lipid transport'})
MERGE (tau_prot:Protein {name: 'Tau', uniprot: 'P10636', function: 'Microtubule binding'})
MERGE (abeta_prot:Protein {name: 'Amyloid-beta', uniprot: 'P05067', function: 'Peptide fragment'})

// Gene codes for Protein
MERGE (trem2_gene)-[:CODES_FOR]->(trem2_prot)
MERGE (apoe_gene)-[:CODES_FOR]->(apoe_prot)
MERGE (mapt_gene)-[:CODES_FOR]->(tau_prot)
MERGE (app_gene)-[:CODES_FOR]->(abeta_prot)

// ============================================
// MERGE SNPs (Clinical Variants from ClinVar)
// ============================================

MERGE (rs75932628:SNP {
  rsid: 'rs75932628',
  variant: 'R47H',
  clinical_significance: 'Pathogenic',
  review_status: '3-star',
  allele_freq: 0.002
})
MERGE (rs429358:SNP {
  rsid: 'rs429358',
  variant: 'ε4 allele',
  clinical_significance: 'Risk factor',
  review_status: '4-star',
  allele_freq: 0.14
})
MERGE (rs63750847:SNP {
  rsid: 'rs63750847',
  variant: 'A152T',
  clinical_significance: 'Likely pathogenic',
  review_status: '2-star',
  allele_freq: 0.0001
})

// SNP-Gene associations
MERGE (rs75932628)-[:ASSOCIATED_WITH]->(trem2_gene)
MERGE (rs429358)-[:ASSOCIATED_WITH]->(apoe_gene)
MERGE (rs63750847)-[:ASSOCIATED_WITH]->(app_gene)

// ============================================
// MERGE KEYWORDS (Research Themes)
// ============================================

MERGE (neuroinflamm:Keyword {term: 'Neuroinflammation'})
MERGE (microglia:Keyword {term: 'Microglia'})
MERGE (geneticRisk:Keyword {term: 'Genetic Risk Factors'})
MERGE (proteinAgg:Keyword {term: 'Protein Aggregation'})
MERGE (immunotherapy:Keyword {term: 'Immunotherapy'})

// Keyword hierarchies
MERGE (microglia)-[:IS_A]->(neuroinflamm)
MERGE (immunotherapy)-[:IS_A]->(neuroinflamm)

// ============================================
// PUBLICATION 1: TREM2 Discovery Paper
// ============================================

MERGE (pub1:Publication {
  pmid: 'PMID35123456',
  title: 'TREM2 variants confer risk for Alzheimers disease through microglial dysfunction',
  year: 2023,
  citations: 342,
  abstract: 'Rare variants in TREM2 increase Alzheimers disease risk 3-fold...'
})
MERGE (pub1)-[:PUBLISHED_IN]->(natGenet)
MERGE (pub1)-[:HAS_AUTHOR]->(chen)
MERGE (pub1)-[:HAS_AUTHOR]->(rodriguez)
MERGE (pub1)-[:HAS_KEYWORD]->(neuroinflamm)
MERGE (pub1)-[:HAS_KEYWORD]->(microglia)
MERGE (pub1)-[:HAS_KEYWORD]->(geneticRisk)
MERGE (pub1)-[:MENTIONS]->(trem2_gene)
MERGE (pub1)-[:MENTIONS]->(trem2_prot)
MERGE (pub1)-[:MENTIONS]->(ad)
MERGE (pub1)-[:MENTIONS_VARIANT]->(rs75932628)

// ============================================
// PUBLICATION 2: APOE Review Paper
// ============================================

MERGE (pub2:Publication {
  pmid: 'PMID35234567',
  title: 'APOE4: The most significant genetic risk factor for late-onset Alzheimers',
  year: 2023,
  citations: 589,
  abstract: 'The APOE ε4 allele increases AD risk in dose-dependent manner...'
})
MERGE (pub2)-[:PUBLISHED_IN]->(neuron)
MERGE (pub2)-[:HAS_AUTHOR]->(williams)
MERGE (pub2)-[:HAS_KEYWORD]->(geneticRisk)
MERGE (pub2)-[:MENTIONS]->(apoe_gene)
MERGE (pub2)-[:MENTIONS]->(apoe_prot)
MERGE (pub2)-[:MENTIONS]->(ad)
MERGE (pub2)-[:MENTIONS_VARIANT]->(rs429358)
MERGE (pub2)-[:CITES]->(pub1)

// ============================================
// PUBLICATION 3: TREM2 Therapeutic Paper (Industry)
// ============================================

MERGE (pub3:Publication {
  pmid: 'PMID35345678',
  title: 'Therapeutic targeting of TREM2 in neurodegenerative diseases',
  year: 2024,
  citations: 127,
  abstract: 'TREM2 agonists show promise in preclinical models...'
})
MERGE (pub3)-[:PUBLISHED_IN]->(cell)
MERGE (pub3)-[:HAS_AUTHOR]->(tanaka)
MERGE (pub3)-[:HAS_AUTHOR]->(rodriguez)
MERGE (pub3)-[:HAS_KEYWORD]->(immunotherapy)
MERGE (pub3)-[:HAS_KEYWORD]->(microglia)
MERGE (pub3)-[:MENTIONS]->(trem2_gene)
MERGE (pub3)-[:MENTIONS]->(trem2_prot)
MERGE (pub3)-[:MENTIONS]->(ad)
MERGE (pub3)-[:MENTIONS]->(pd)
MERGE (pub3)-[:CITES]->(pub1)

// ============================================
// PUBLICATION 4: Multi-omics Paper
// ============================================

MERGE (pub4:Publication {
  pmid: 'PMID35456789',
  title: 'Integrated genomics reveals novel Alzheimers disease susceptibility loci',
  year: 2024,
  citations: 234,
  abstract: 'Genome-wide association analysis identifies BIN1 and CD33...'
})
MERGE (pub4)-[:PUBLISHED_IN]->(nature)
MERGE (pub4)-[:HAS_AUTHOR]->(schmidt)
MERGE (pub4)-[:HAS_AUTHOR]->(kumar)
MERGE (pub4)-[:HAS_AUTHOR]->(chen)
MERGE (pub4)-[:HAS_KEYWORD]->(geneticRisk)
MERGE (pub4)-[:MENTIONS]->(bin1_gene)
MERGE (pub4)-[:MENTIONS]->(trem2_gene)
MERGE (pub4)-[:MENTIONS]->(apoe_gene)
MERGE (pub4)-[:MENTIONS]->(ad)

// ============================================
// PUBLICATION 5: Tau/FTD Paper
// ============================================

MERGE (pub5:Publication {
  pmid: 'PMID35567890',
  title: 'MAPT mutations in frontotemporal dementia and Alzheimers overlap',
  year: 2023,
  citations: 178,
  abstract: 'Tau protein dysfunction links multiple neurodegenerative diseases...'
})
MERGE (pub5)-[:PUBLISHED_IN]->(neuron)
MERGE (pub5)-[:HAS_AUTHOR]->(kumar)
MERGE (pub5)-[:HAS_KEYWORD]->(proteinAgg)
MERGE (pub5)-[:MENTIONS]->(mapt_gene)
MERGE (pub5)-[:MENTIONS]->(tau_prot)
MERGE (pub5)-[:MENTIONS]->(ftd)
MERGE (pub5)-[:MENTIONS]->(ad)

// ============================================
// PUBLICATION 6: Industry-Academic Collaboration
// ============================================

MERGE (pub6:Publication {
  pmid: 'PMID35678901',
  title: 'Clinical validation of TREM2 R47H variant in diverse populations',
  year: 2024,
  citations: 95,
  abstract: 'Multi-center study confirms TREM2 variant pathogenicity...'
})
MERGE (pub6)-[:PUBLISHED_IN]->(natGenet)
MERGE (pub6)-[:HAS_AUTHOR]->(tanaka)
MERGE (pub6)-[:HAS_AUTHOR]->(williams)
MERGE (pub6)-[:HAS_AUTHOR]->(schmidt)
MERGE (pub6)-[:HAS_KEYWORD]->(geneticRisk)
MERGE (pub6)-[:MENTIONS]->(trem2_gene)
MERGE (pub6)-[:MENTIONS]->(ad)
MERGE (pub6)-[:MENTIONS_VARIANT]->(rs75932628)
MERGE (pub6)-[:CITES]->(pub1)
MERGE (pub6)-[:CITES]->(pub3)

Cypher Queries

The following example queries demonstrate how publication intelligence can be extracted from the graph.

Institutional Hotspots

All of the below queries are focused on finding hotspots of research activity which can be useful in a number of competitive intelligence scenarios.

Which institutions are publishing on a drug target and a specific disease?

You might use this to identify leading research centers for potential collaborations, or to monitor competitor activity in a therapeutic area.

WITH
  "TREM2" AS targetGene,
  "Alzheimer's Disease" AS targetDisease
MATCH
  (pub:Publication)-[:MENTIONS]->(gene:Gene {symbol: targetGene}),
  (pub)-[:MENTIONS]->(disease:Disease {name: targetDisease}),
  (pub)-[:HAS_AUTHOR]->(author:Author)-[:AFFILIATED_WITH]->(inst:Institution)
RETURN
  inst.name AS Institution,
  inst.type AS Type,
  count(DISTINCT pub) AS Publications,
  collect(DISTINCT author.name) AS Researchers,
  avg(pub.citations) AS AvgCitations
ORDER BY Publications DESC, AvgCitations DESC;

Which institutions collaborate most frequently on publications for a target and disease?

This is useful for identifying strong institutional partnerships and collaboration networks in a therapeutic area, as well as potential gaps where new collaborations could be fostered.

WITH
  "TREM2" AS targetGene,
  "Alzheimer's Disease" AS targetDisease
MATCH
  (pub:Publication)-[:MENTIONS]->(gene:Gene {symbol: targetGene}),
  (pub)-[:MENTIONS]->(disease:Disease {name: targetDisease}),
  (pub)-[:HAS_AUTHOR]->(a1:Author)-[:AFFILIATED_WITH]->(inst1:Institution),
  (pub)-[:HAS_AUTHOR]->(a2:Author)-[:AFFILIATED_WITH]->(inst2:Institution)
WHERE inst1.name < inst2.name
RETURN
  inst1.name AS Institution1,
  inst2.name AS Institution2,
  count(DISTINCT pub) AS SharedPublications,
  collect(DISTINCT pub.pmid) AS PMIDs
ORDER BY SharedPublications DESC, Institution1, Institution2

Which authors are publishing on a drug target and a specific disease?

You might use this query to identify key opinion leaders in a therapeutic area, or to monitor competitor researchers working on specific targets.

The end result might be that you hire some of these researchers as consultants or advisors, or you might reach out to them for potential collaborations.

WITH
  "TREM2" AS targetGene,
  "Alzheimer's Disease" AS targetDisease
MATCH
  (pub:Publication)-[:MENTIONS]->(gene:Gene {symbol: targetGene}),
  (pub)-[:MENTIONS]->(disease:Disease {name: targetDisease}),
  (pub)-[:HAS_AUTHOR]->(author:Author)
RETURN
  author.name AS Author,
  author.h_index AS `H-Index`,
  count(DISTINCT pub) AS Publications
ORDER BY Publications DESC, `H-Index` DESC;

Which keywords are associated with a drug target and a specific disease?

This would give you a broad sense of the main research themes and trends in the literature around that target and disease, helping you identify emerging areas of interest or gaps in the research landscape.

WITH
  "TREM2" AS targetGene,
  "Alzheimer's Disease" AS targetDisease
MATCH
  (pub:Publication)-[:MENTIONS]->(gene:Gene {symbol: targetGene}),
  (pub)-[:MENTIONS]->(disease:Disease {name: targetDisease}),
  (pub)-[:HAS_KEYWORD]->(keyword:Keyword)
RETURN
  keyword.term AS Keyword,
  count(DISTINCT pub) AS Publications
ORDER BY Publications DESC;

Clinical Validation

Clinical validation of gene-disease associations allows you to prioritize targets with strong evidence.

Find emerging gene-disease associations through clinical variants

This query identifies genes linked to diseases via clinically significant variants (Single Nucleotide Polymorphisms) mentioned in publications, helping prioritize targets with strong clinical evidence.

WITH ['Pathogenic', 'Likely pathogenic'] AS significantClasses
MATCH
  (snp:SNP)-[:ASSOCIATED_WITH]->(gene:Gene),
  (pub:Publication)-[:MENTIONS_VARIANT]->(snp),
  (pub)-[:MENTIONS]->(disease:Disease)
WHERE snp.clinical_significance IN significantClasses
RETURN
  gene.symbol AS Gene,
  gene.name AS GeneName,
  snp.rsid AS ClinicalVariant,
  snp.variant AS Mutation,
  disease.name AS Disease,
  count(DISTINCT pub) AS Publications,
  snp.clinical_significance AS ClinicalSignificance
ORDER BY Publications DESC;

Find publications that mention a clinical variant and a disease.

This is looking for publications that mention both a clinical variant (SNP rsID) and a disease, helping to identify studies that provide clinical context for genetic findings.

MATCH
  (snp:SNP)-[:ASSOCIATED_WITH]->(gene:Gene),
  (pub:Publication)-[:MENTIONS_VARIANT]->(snp),
  (pub)-[:MENTIONS]->(disease:Disease)
RETURN
  pub.title AS Publications,
  pub.year AS Year,
  snp.rsid AS ClinicalVariant,
  snp.variant AS Mutation,
  disease.name AS Disease
ORDER BY Publications DESC;

Collaboration Networks

Research collaborations between academic institutions and industry can accelerate drug discovery, and can be used to identify potential partners.

Map research collaboration networks

We’re looking for pairs of institutions (one academic, one industry) that have co-authored publications together, along with the diseases they are studying collaboratively, which can help identify strong partnerships and potential collaboration opportunities.

MATCH
  (auth1:Author)-[:AFFILIATED_WITH]->(inst1:Institution),
  (auth2:Author)-[:AFFILIATED_WITH]->(inst2:Institution),
  (pub:Publication)-[:HAS_AUTHOR]->(auth1),
  (pub)-[:HAS_AUTHOR]->(auth2)
WHERE
  inst1.name < inst2.name
  AND inst1.type <> inst2.type  // Cross-sector collaboration
MATCH (pub)-[:MENTIONS]->(disease:Disease)
RETURN
  inst1.name AS AcademicInstitution,
  inst2.name AS IndustryPartner,
  count(DISTINCT pub) AS CollaborativePublications,
  collect(DISTINCT auth1.name + ' & ' + auth2.name) AS ResearcherPairs,
  collect(DISTINCT disease.name) AS DiseasesStudied
ORDER BY CollaborativePublications DESC;

Key Opinion Leaders

Identify the most influential researchers in a specific therapeutic area. These might be people you want to collaborate with or potentially aim to hire.

WITH
  "TREM2" AS targetGene,
  "Alzheimer's Disease" AS targetDisease
MATCH
  (pub:Publication)-[:MENTIONS]->(gene:Gene {symbol: targetGene}),
  (pub)-[:MENTIONS]->(disease:Disease {name: targetDisease}),
  (pub)-[:HAS_AUTHOR]->(author:Author)
RETURN
  author.name AS Author,
  author.h_index AS `H-Index`,
  count(DISTINCT pub) AS Publications
ORDER BY Publications DESC, `H-Index` DESC;

Find "bridge" researchers connecting different research themes

Identify key opinion leaders who publish across multiple domains.

MATCH (author:Author)<-[:HAS_AUTHOR]-(pub:Publication)-[:HAS_KEYWORD]->(keyword:Keyword)
WITH
  author,
  collect(DISTINCT keyword.term) AS themes,
  count(DISTINCT pub) AS pubCount
WHERE size(themes) >= 2
MATCH (author)-[:AFFILIATED_WITH]->(inst:Institution)
RETURN
  author.name AS Researcher,
  inst.name AS Institution,
  author.h_index AS `H-Index`,
  themes AS ResearchThemes,
  pubCount AS Publications
ORDER BY size(themes) DESC, author.h_index DESC;

Multi-target Research Trends

Discover multi-target research trends (polygenic approaches) in diseases.

Find publications linking multiple genes to the same disease (systems biology approach).

WITH "Alzheimer's Disease" AS targetDisease
MATCH
  (pub:Publication)-[:MENTIONS]->(disease:Disease {name: targetDisease}),
  (pub)-[:MENTIONS]->(gene:Gene),
  (pub)-[:HAS_AUTHOR]->(author:Author)-[:AFFILIATED_WITH]->(inst:Institution)
WITH
  pub,
  disease,
  collect(DISTINCT gene.symbol) AS genes,
  inst
WHERE size(genes) >= 2  // Publications mentioning 2+ genes
RETURN
  pub.title AS Publication,
  pub.pmid AS PMID,
  pub.year AS Year,
  inst.name AS Institution,
  disease.name AS Disease,
  genes AS GenesStudied,
  size(genes) AS NumberOfGenes,
  pub.citations AS Citations
ORDER BY NumberOfGenes DESC, Citations DESC

Publication-based competitive intelligence (CI) can be powerfully connected to patent-based CI by leveraging overlapping entities within both datasets. Scientific publications often provide the foundational knowledge that patents build upon, with many patents explicitly citing key publications during the patent application process. Entities such as genes, drug targets, and molecular variants commonly appear in both scholarly articles and patent claims, allowing for the mapping of innovation trajectories from basic research to intellectual property. Furthermore, authors of high-impact publications may also appear as inventors or advisors in patents, making them key opinion leaders and potential collaboration partners. Integrating publication and patent CI enables organizations to identify emerging hot spots, influential contributors, and the translation of novel scientific discoveries into protectable inventions.

See also: Patent Intelligence