pathXcite: Enhancing Article Retrieval

What you'll learn

Convert a research goal into a minimal set of must-have & nice-to-have terms.
Compose robust PubMed queries using phrases, Boolean logic, field tags, and MeSH.
Apply precise filters (article type, organism, date, text availability) to shape the evidence.
Create reusable query templates for common goals (e.g., pathway focus, biomarkers, therapy names).

Prereq: You have a project open. Run queries in the app's Web Browser and use Scan For Articles to collect IDs when ready.

Core principles

1) Start from the goal

Frame your intent in three facets:

Topic: disease/phenotype/condition.
Mechanism: pathway/process/gene family/drug class.
Context: organism, tissue, model, article type, date.

Decide which are must, which are nice, and what to exclude.

2) Prefer precision, then widen

Use phrases with quotes: "Alzheimer disease".
Bind terms to fields: [tiab], [mh], [pt], [dp].
Mix AND/OR/NOT to shape sets; add synonyms with OR.
Layer filters (article type, species, date) to keep noise low.

Field tags (PubMed): [tiab] title/abstract; [mh] MeSH term; [pt] publication type [dp] date; [la] language; [jour] journal (name/abbrev)

Goal → query recipes

Pathway-specific evidence (avoid generic background)

("Alzheimer Disease"[mh] OR "Alzheimer disease"[tiab])
AND ("Synaptic Transmission"[mh] OR "synaptic plasticity"[tiab] OR "LTP"[tiab] OR "long-term potentiation"[tiab])
AND (humans[mh] OR mice[mh])
NOT review[pt]
AND 2019:3000[dp]

Biomarkers (diagnostic/prognostic) with human focus

("Alzheimer disease"[tiab] OR "Alzheimer Disease"[mh])
AND (biomarker*[tiab] OR "ROC curve"[tiab] OR sensitivity[tiab] OR specificity[tiab])
AND humans[mh]
AND (case-control[pt] OR cohort[pt] OR clinical trial[pt])
AND 2021:3000[dp]

Drug/therapy mentions tied to mechanism

("Alzheimer disease"[tiab] OR "Alzheimer Disease"[mh])
AND (donepezil[tiab] OR lecanemab[tiab] OR "anti-amyloid"[tiab])
AND ("Amyloid beta-Peptides"[mh] OR "amyloid beta"[tiab] OR "tau protein"[tiab])
NOT review[pt]
AND english[la]

Gene-family focus (synonyms included)

("Alzheimer disease"[tiab] OR "Alzheimer Disease"[mh])
AND (APOE[tiab] OR "apolipoprotein E"[tiab] OR CLU[tiab] OR "clusterin"[tiab] OR BIN1[tiab])
AND (expression[tiab] OR upregulat*[tiab] OR differential[tiab])
AND (brain[tiab] OR hippocampus[tiab] OR cortex[tiab])

Keep reviews on purpose (for quick background)

("Alzheimer disease"[tiab]) AND (synapse*[tiab] OR microglia[tiab]) AND review[pt] AND 2022:3000[dp]

Replace disease/mechanism/gene terms with your own; add synonyms with OR. Tighten or loosen with date ([dp]) and type ([pt]) as needed.

Filters that matter

Keep

Article type: clinical trial[pt], case-control[pt], cohort[pt], systematic review[pt]
Organism: humans[mh], mice[mh], rats[mh]
Date window: 2020:3000[dp] (or past 5-7y for recency)
Language: english[la] (if helpful)
Text availability: “Free full text” (PubMed filter UI)

Avoid

Over-broad NOTs (risk of removing relevant items unintentionally).
Unquoted phrases for multi-word concepts (e.g., write "long-term potentiation").
Unscoped acronyms (e.g., LTP) without a spelled-out synonym.

MeSH specifics: Use [mh:noexp] to prevent exploding to child terms when you need a narrow slice.

Reusable templates

General template

(
  ""[tiab] OR ""[mh]
) AND (
  ""[mh] OR ""[tiab] OR [tiab] OR [tiab]
) AND (
  humans[mh] OR mice[mh]
)
AND :3000[dp]
NOT review[pt]

Diagnostic/prognostic focus

(""[tiab] OR ""[mh])
AND (biomarker*[tiab] OR "diagnos*"[tiab] OR "prognos*"[tiab] OR "ROC"[tiab])
AND (case-control[pt] OR cohort[pt] OR clinical trial[pt])

Intervention/mechanism mapping

(""[tiab])
AND ([tiab] OR [tiab])
AND (""[mh] OR ""[tiab])
NOT review[pt]

Using these in pathXcite

Open the Web Browser module and paste a crafted query into PubMed's search bar.
Apply PubMed UI filters (article type, text availability, dates, species) to narrow results.
Use Scan For Articles to capture IDs from the current result page(s).
Proceed when you have a set that matches your goal (you can refine later in Document Insights).

You can iterate: adjust the query/filters, rescan, and compare sets downstream.

Common pitfalls & fixes

Too broad? Add [tiab] to key terms, require a mechanism term, or add a date window.
Too narrow? Remove NOT review[pt], broaden synonyms, switch from [mh:noexp] to [mh].
Acronym collisions? Pair acronyms with spelled-out phrases via OR.
Background-heavy? Keep reviews for orientation, but run a second query excluding review[pt] for results-driven evidence.

Pocket reference

Operators

AND intersection; OR synonyms; NOT exclude
Quotes for phrases: "microglial activation"

Fields

[tiab] Title/Abstract; [mh] MeSH; [pt] Type
[dp] Date; [la] Language; [jour] Journal

Next steps

Continue with: Analysing the section plot Understanding GF-IDF & Count