← Back to search

How It Works

A researcher's guide to every feature in LOOM CRISPR Search.

New to CRISPR or pathogen genomics? Start with Learn for a visual, step-by-step introduction before diving into this feature reference.

1. What is this tool?

LOOM CRISPR Search is an open-science platform for discovering CRISPR-based diagnostic targets across pathogen genomes. It uses a 195 KB WebAssembly binary (built on a Burrows-Wheeler Transform FM-index) to perform sub-millisecond exact-match searches across millions of genome sequences — entirely in your browser. No data leaves your machine.

The database contains 140,000 pre-computed CRISPR targets across 13 human pathogens extracted from over 3 million genomes downloaded from NCBI.

2. Pathogen Cards & Diagnostic Priority

The landing page shows a card for each pathogen. Each card displays the number of targets and genomes indexed.

Diagnostic Priority Badge

Many cards have a colored badge in the top-right corner:

The score (0–100) is a composite of five weighted factors:

Tip: Hover over any badge to see the exact numeric score and the factors behind it.

Top Diagnostic Opportunities

The three featured cards at the top highlight pathogens with the highest priority scores. These include a brief summary (e.g., "No CRISPR dx published · 21,000–143,000 deaths/yr") so you can immediately see why they rank high.

3. Targets Table — Columns Explained

After selecting a pathogen, the table shows one row per CRISPR target site (23-mer). Here's what each column means:

Sequence (23-mer)

The 23-nucleotide CRISPR target sequence, shown with color-coded bases (A=green, T=red, G=gold, C=blue). This includes the 20-nt guide + 3-nt PAM motif. Badges may appear next to the sequence (see Cross-Reactivity and Resistance sections below).

Gene

The gene the target falls within, or intergenic if it's between genes. Click on a gene tag to jump to the Disease Context tab and highlight that gene in the gene map.

Position

The start position (0-based offset) within the indexed genome.

PAM

The protospacer adjacent motif. NGG = SpCas9-compatible (most widely used). TTTN = Cas12a-compatible. = no canonical PAM detected (still a valid target site for PAM-less Cas variants).

Coverage

How many of the indexed genomes contain this exact 23-mer sequence. Shown as both a count and a percentage. Example: 1,832 67.6% means 1,832 out of 2,711 Zika genomes have this sequence.

Higher coverage = more conserved target = works across more strains = better diagnostic candidate.

Score

Guide RNA quality score (see next section).

Actions

BLAST link (see BLAST section below).

4. Guide RNA Quality Score

Each guide gets a composite quality score from 0 to 100, displayed as a colored circle:

The score is computed from six factors:

Tip: Hover over any score badge to see the individual GC%, seed-region GC%, and max homopolymer run length.

5. Cross-Reactivity Badges

For the top 100 NGG guides per pathogen, we ran exact-match searches against 7 animal/human host genomes (human GRCh38, pig, bat, chicken, cow, camel, mouse) to check whether the guide sequence also appears in a host genome — which would cause off-target effects in a diagnostic assay.

99.7% of the tested guides are specific. Guides without a badge were not among the top 100 NGG guides for that pathogen and have not been tested.

6. BLAST Links

Every row has a BLAST link in the Actions column. Clicking it opens NCBI BLAST (blastn) in a new tab with the guide sequence pre-filled and the database set to nt (all NCBI nucleotide sequences).

You need to click the blue "BLAST" button on the NCBI page to run the search — NCBI does not allow automated submission via URL because each search uses compute resources on their servers.

What BLAST tells you: Which organisms and genomic regions contain this exact (or similar) sequence. Use it to verify specificity beyond our pre-computed host checks, or to explore evolutionary conservation.

7. Drug-Resistance Overlay

Guides that overlap known drug-resistance mutation regions show a ☢ geneSymbol badge next to the sequence. This means the target site sits within a genomic region where drug-resistance mutations are known to occur.

Why it matters: Targeting resistance regions can be a double-edged sword — it could detect resistant strains specifically, but mutations in that region may also cause the guide to lose binding in resistant variants.

Hover over the badge to see the specific drug and mutation region name.

Coverage by pathogen:

8. Research Novelty Filter

The dropdown filter lets you view:

Genes are identified using NCBI gene annotations and matched against PubMed-indexed publications via ontology-bridged symbol resolution.

9. Live Search

The Live Search tab lets you search for any arbitrary DNA sequence against a pathogen's FM-index — loaded directly into your browser as a WASM binary.

  1. Select a pathogen from the cards.
  2. Go to the Live Search tab.
  3. Click "Load Index" (downloads the FM-index, typically 5–200 MB).
  4. Type or paste any DNA sequence. Results appear instantly (< 1 ms).

This searches the full indexed genome, not just pre-computed targets. You can search for primers, probes, or any sequence of interest.

Note: Very large indexes (e.g., Human GRCh38 at 1.5 GB) may take longer to download but still search in under 1 ms once loaded.

10. Panel Designer

The Panel Designer tab helps you design a minimal multiplex diagnostic panel — a small set of CRISPR guides that can distinguish between multiple pathogens in a single assay.

  1. Select which pathogens to include (or use a preset: Respiratory, Hemorrhagic, STI, All).
  2. Set minimum conservation threshold (default: 70%) and minimum quality score (default: 50).
  3. Click Design Panel.

The algorithm uses a greedy set-cover approach: for each pathogen, it picks the highest-scoring guide that is unique to that pathogen (not found as a top candidate in others). The result is a table showing one distinguishing guide per pathogen.

Click Export Panel CSV to download the panel as a spreadsheet for ordering oligos or sharing with your lab.

If a pathogen shows "no qualifying guides": Relax the conservation or score thresholds. Pathogens with very large genome diversity (e.g., Dengue with 55,000 genomes) may not have guides reaching 70% conservation.

11. Disease Context Tab

After selecting a pathogen, the Disease Context tab shows biomedical ontology data:

12. Exports

Two export buttons are available when viewing a pathogen's targets:

The Panel Designer also has its own Export Panel CSV button for the designed panel.

13. Methods & Data Sources

Full methodology details are available in the Data Sources tab within the app. Key points:

LOOM CRISPR Search — Open-science CRISPR target discovery

Alvaro Videla Godoy — 2026