🔬 BioKhoj Help

BioKhoj (from Hindi khoj, meaning "search" or "discovery") is a personal literature radar for biomedical researchers. It continuously monitors the published literature for topics you care about and surfaces the most relevant new papers using an intelligent Signal Score.

1. Add entities to your watchlist — genes (BRCA1), diseases (glioblastoma), pathways (mTOR signaling), drugs (olaparib), or any biomedical concept
2. BioKhoj finds papers — queries PubMed, bioRxiv, and OpenAlex for recent publications mentioning your watchlist entities
3. Papers are scored — each paper gets a Signal Score (0–100) based on recency, citation velocity, journal tier, entity match strength, and novelty of co-mentions
4. You stay informed — high-signal papers surface to the top, co-mention alerts flag unexpected connections, and trend charts show how your topics evolve over time

When you first open BioKhoj, you will see an empty watchlist. Here is how to get started in 60 seconds:

1. Click "+ Add Entity" and type TP53 — BioKhoj auto-classifies it as a gene
2. Add a second entity: glioblastoma — classified as a disease
3. Click "Refresh Feed" — BioKhoj fetches recent papers mentioning either entity
4. Papers appear sorted by Signal Score — the most relevant papers are at the top
5. Look for co-mention badges — these flag papers that mention both TP53 and glioblastoma together, which may indicate novel connections

Your watchlist is the core of BioKhoj. Each entry is a biomedical concept you want to track.

• Add — click "+ Add Entity", type a name, and press Enter. BioKhoj auto-classifies the entity type (gene, disease, pathway, drug, organism, method, or custom).
• Remove — click the × button next to any entity to remove it from your watchlist
• Edit — click an entity name to edit it inline, or change its type/tags
• Bulk add — paste a comma-separated list of entities to add multiple at once

Each watchlist entity can be assigned a priority level that affects how papers are scored:

Use high priority for your primary research focus and low priority for tangentially related topics you want to keep an eye on.

Tag entities with custom labels (e.g., thesis, grant-R01, collab-smith-lab) to organize your watchlist. Tags are used for:

• Filtering — filter the paper feed to show only papers matching entities with a specific tag
• Grouping — view your watchlist grouped by tag in the sidebar
• Export — export papers filtered by tag

Pause — temporarily pause an entity to exclude it from feed queries without removing it. Paused entities appear dimmed in the watchlist and can be resumed at any time.

Auto-classification — when you add an entity, BioKhoj automatically classifies it into one of these types:

You can override the auto-classification by clicking the type badge and selecting a different type.

BioKhoj queries multiple sources to find papers related to your watchlist entities:

• PubMed — via NCBI E-utilities (esearch + esummary), fetches recent papers matching entity names using field-tagged queries (e.g., BRCA1[Gene] AND human[Organism])
• bioRxiv/medRxiv — via the bioRxiv API, fetches recent preprints
• OpenAlex — via the OpenAlex API, provides citation data, author information, and concept tagging

Results are deduplicated by DOI/PMID, merged, and scored. The feed updates when you click Refresh or on the schedule configured in Settings.

Every paper in the feed shows a Signal Score badge indicating its relevance to you:

Click the badge to see the full score breakdown for that paper.

• Sort by — Signal Score (default), date, citation count, or journal tier
• Filter by entity — show papers matching a specific watchlist entity
• Filter by tag — show papers matching entities with a specific tag
• Filter by score — show only high-signal papers (>=70) or hide low-signal (<40)
• Filter by source — PubMed only, preprints only, or all
• Time range — last 7 days, 30 days, 90 days, or 1 year

The feed toolbar is the primary control row above the paper list. It contains three filter toggle buttons and an Insights dropdown:

• Multi-entity — show only papers mentioning 2+ watched entities (co-mention filter)
• Unread — show only papers you have not yet opened or marked as read
• High signal — show only papers scoring above the high-signal threshold (default ≥70)

All three filters are composable — enable any combination to narrow your feed. The "Insights" dropdown (right side of the toolbar) provides four actions:

• Journal Club — curated list of the week's top papers for group discussion
• Collaborators — papers by watched authors or co-authors of your saved papers
• Weekly Digest — formatted summary of the past week's high-signal papers
• Share watchlist — copy a shareable URL encoding your watchlist entities

Two dedicated toggle buttons in the feed toolbar let you quickly focus on what matters:

• Unread — filters the feed to papers you have not opened. BioKhoj tracks read status automatically when you click into a paper or expand its abstract. The feed subtitle shows the count of unread papers.
• High Signal — filters to papers scoring at or above the high-signal threshold (default ≥70). Combine with Unread to see only unread high-signal papers — the fastest way to triage new results.

Both filters compose with the Multi-entity co-mention filter, so you can apply all three simultaneously to find unread, high-signal, cross-topic papers.

Each paper card in the feed shows a 180-character abstract preview followed by a "more" link. Click "more" to expand the full abstract inline beneath the paper title. Click "less" to collapse it back to the preview. This replaces the old hover tooltip and works better on touch devices and narrow screens.

On desktop, the sidebar displays your watchlist entities with colored activity dots that reflect real-time feed state:

The dots update after each feed refresh. The purple pulse animation draws attention to entities with the most actionable new results.

The Signal Score is a composite score from 0 to 100 that ranks papers by how relevant and important they are to your research. It is computed from six weighted components:

You can customize these thresholds in Settings.

The Signal Mute Threshold is a slider in Settings (range 0–80) that hides all papers scoring below your chosen cutoff from the feed. Unlike the high-signal filter toggle (which shows only ≥70), the mute threshold is a persistent baseline — papers below it never appear in your feed at all.

When a mute threshold is active, the feed subtitle shows the active value (e.g., "Muted below 30"). Set it to 0 to see everything, or raise it to 20–40 to suppress noise from tangential matches. The threshold does not affect papers already saved to your reading list.

BioKhoj can suggest related concepts to expand your watchlist coverage. When enabled, it uses:

• OpenAlex concepts — papers matching your entities are tagged with OpenAlex concepts; frequently co-occurring concepts that are not already in your watchlist are suggested
• MeSH term expansion — for disease and pathway entities, BioKhoj looks up the MeSH hierarchy and suggests parent/child terms (e.g., if you watch "glioblastoma", it may suggest "glioma" or "brain neoplasms")
• Gene interaction partners — for gene entities, suggests commonly co-studied genes based on STRING interaction data

Suggestions appear in a "Suggested" section below your watchlist with a brief rationale for each.

Each suggestion has an "+ Add" button. Click it to instantly add the concept to your watchlist with auto-classification. Click "Dismiss" to hide the suggestion permanently.

A co-mention alert fires when a paper discusses two or more of your watchlist entities together in a way that is statistically unusual based on historical co-occurrence data.

For example, if your watchlist contains BRCA1 and ferroptosis, and a new paper discusses both, BioKhoj flags this as a co-mention alert because these two concepts rarely appear together in the literature — this could represent a novel research direction.

BioKhoj computes novelty by comparing the observed co-occurrence count against the expected count based on each entity's individual frequency:

• Expected co-occurrence = (freq_A * freq_B) / total_papers
• Observed co-occurrence = actual number of papers mentioning both
• If observed is significantly lower than expected (rare combination), the co-mention novelty score is high

Co-mention alerts appear with a CO-MENTION badge in the feed and contribute to the Signal Score's co-mention novelty component.

The Trends view shows publication volume over time for each watchlist entity:

• Sparklines — compact inline charts next to each entity in the watchlist showing 12-month publication trend at a glance
• Full trend chart — click any entity to see a detailed chart with monthly publication counts, citation averages, and preprint vs. published ratios
• Overlay mode — select multiple entities to overlay their trends on a single chart for comparison

• 3 months — recent activity, weekly granularity
• 6 months — medium-term view, bi-weekly granularity
• 1 year — annual trend, monthly granularity (default)
• 5 years — long-term trajectory, quarterly granularity

Trend data is fetched from OpenAlex and cached locally for 24 hours.

Above the trend charts, BioKhoj displays a text-based insights summary box with four data points:

• Entity activity changes — week-over-week publication volume shifts (e.g., "TP53 up 32% this week", "CRISPR down 12%")
• Top journals — which journals published the most papers matching your watchlist this period
• Co-mention spikes — entity pairs whose co-occurrence rate spiked above historical baseline
• High-signal count — total number of papers scoring ≥70 in the current period

The summary refreshes whenever trend data is recalculated and provides a quick narrative you can scan before diving into the charts.

Save any paper from the feed to your Reading List for later reading. Your reading list is stored locally and persists across sessions.

• Save — click the bookmark icon on any paper in the feed
• Notes — add personal notes to saved papers (e.g., "Check methods section", "Relevant to aim 2")
• Tags — tag saved papers with custom labels for organization
• Status — mark papers as Unread, Reading, Read, or Archived

• Sort by — date saved, Signal Score, publication date, or title
• Filter by — status (unread/reading/read), tags, or search by title/author
• Export — export your reading list as BibTeX, RIS, Markdown, CSV, pandas DataFrame, or R tibble (see Export and Export to R/Python sections)

Add specific authors to your watch list to get notified when they publish new papers:

• Add from paper — click an author name in any paper's detail view, then click "Watch Author"
• Add manually — enter an author name or ORCID in the Author Tracking section
• OpenAlex matching — BioKhoj resolves author names to OpenAlex author IDs for reliable tracking across name variants

When a watched author publishes a new paper, it appears in your feed with an WATCHED AUTHOR badge and receives a +7 bonus to the Author Reputation component of its Signal Score. If notifications are enabled, you will get a desktop notification.

• Current feed — export all papers currently visible in the feed (respects active filters)
• Reading list — export saved papers (with notes and tags)
• Selected papers — export only checked papers
• Watchlist — export your entity watchlist as JSON for backup or sharing

If you use BioGist (the entity scanner extension), you can import detected entities directly into your BioKhoj watchlist:

1. Scan a paper in BioGist to detect genes, drugs, pathways, etc.
2. Click "Send to BioKhoj" in BioGist's sidebar (or use the export menu)
3. BioKhoj receives the entities and adds them to your watchlist with auto-classification

Both extensions use a shared entity format, so entities transfer cleanly between them. Pinned entities in BioGist are imported as high-priority watchlist entries.

• 100% local storage — your watchlist, reading list, notes, tags, and settings are stored in your browser's local storage. Nothing is uploaded to any server.
• API calls for paper queries only — the only network requests BioKhoj makes are to PubMed (NCBI), bioRxiv, and OpenAlex to fetch paper metadata. These are standard public APIs.
• No proxy server — all API calls go directly from your browser to the data source. There is no Oric Labs server in between.
• No analytics — no telemetry, tracking, or usage analytics of any kind.
• No account required — BioKhoj works without sign-up or login.
• Exportable data — you can export your entire watchlist and reading list as JSON at any time for backup or portability.

BioKhoj ships with 5 curated starter packs to help you bootstrap a watchlist quickly. Each pack contains 6–8 hand-picked entities relevant to a major research area. Open the Watchlist tab and click a pack to import all its entities in one click. Available packs:

• Cancer Genomics — TP53, BRCA1, KRAS, EGFR, PIK3CA, PTEN, MYC, APC
• CRISPR & Gene Editing — CRISPR-Cas9, base editing, prime editing, guide RNA, off-target effects, Cas12a
• Single-Cell Genomics — scRNA-seq, UMAP, cell atlas, spatial transcriptomics, trajectory inference, Seurat
• Clinical Variants — ClinVar, ACMG, pathogenic variant, VUS, pharmacogenomics, BRCA2
• Metagenomics — 16S rRNA, microbiome, shotgun metagenomics, taxonomic profiling, antimicrobial resistance, QIIME2

Preset packs are auto-hidden once your watchlist has more than 5 entities, keeping the interface clean. You can always re-access them from Settings.

Toggle the "Multi-entity" button in the feed toolbar to filter the feed down to papers that mention 2 or more of your watched entities. This filter composes with the Unread and High Signal toggles — enable all three to find unread, high-signal papers bridging multiple topics. This is exactly the kind of cross-domain work that is easy to miss with single-keyword searches.

Click "Share" in the feed toolbar to copy a shareable URL containing your entire watchlist encoded in the URL hash. Anyone who opens the link will be prompted to import the entities (deduplicated against their existing watchlist). No server is involved — the watchlist data is encoded directly in the URL fragment, so it works entirely client-side.

Hold Shift and click papers in the feed to multi-select them. A "Tag selected" button appears in the toolbar. Click it to apply a tag (e.g., methods to read, thesis chapter 3) to all selected papers at once. This is much faster than tagging papers one by one when triaging a large batch of new results.

On app open, BioKhoj checks whether any paper in your reading list has gained 5 or more new citations since the last check. If so, it shows a toast notification with the paper title and the citation spike count. This helps you spot papers that are suddenly gaining traction in the community — a potential signal that the work is being validated or sparking follow-up research.

The reading list export dropdown now includes three additional formats alongside BibTeX, RIS, and Markdown:

• Copy as pandas — generates a pd.DataFrame(...) snippet ready to paste into a Jupyter notebook
• Copy as R tibble — generates a tibble(...) snippet ready to paste into RStudio
• CSV file export — downloads a standard CSV file with columns for title, authors, DOI, journal, date, Signal Score, and tags

All clipboard formats copy directly to your clipboard — just paste into your analysis environment.

The Settings tab shows a visual progress bar of NCBI API calls used in the last hour. It is color-coded: green when under 50% usage, amber at 50–80%, and red above 80%. The bar also displays your current rate limit (3 req/s without an API key, 10 req/s with one) and the remaining budget. It auto-refreshes every 5 seconds while the Settings tab is open.

When you have been away for more than 1 day, BioKhoj automatically fetches papers you missed. It queries in 7-day windows, going back up to 30 days maximum. A progress bar shows the catch-up status, and a "Welcome back" banner offers three options:

• Review backlog — browse the missed papers sorted by Signal Score
• Mark all read — dismiss the backlog and start fresh
• Dismiss — hide the banner (backlog papers remain in the feed)

For Chrome extension users, BioKhoj adds context menu actions for quick interactions:

• Selected text — right-click highlighted text to "Watch in BioKhoj" (adds to watchlist) or "Lookup in BioKhoj" (opens feed filtered to that term)
• Links to PubMed, bioRxiv, or DOI URLs — right-click a link to "Save to BioKhoj reading list", which fetches the paper metadata and adds it directly to your reading list

BioKhoj — your personal literature radar for biomedical research.

Built by Oric Labs. Part of the BioLang ecosystem.

Report issues or request features: GitHub Issues

BioKhoj is designed for researchers, bioinformaticians, and graduate students who want to stay on top of the literature without manually checking PubMed every day. It brings the most relevant papers to you, scored and ranked by what matters to your specific research interests.