Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI agent for microbiologists

Isolate Relatedness Agent

Give the scientist stable, evidence-backed clusters and a note on which links are uncertain.

Isolate Relatedness Agent: what goes in, what the agent does and what you get

What it does

When a new bacterial isolate arrives, someone has to decide whether it belongs to a known cluster, and results change with the threshold used. This agent compares the new sequences with the existing collection using the lab's method, such as allele differences or SNP distance. It groups isolates into clusters and checks each cluster against dates, places and sources to see whether the epidemiology fits. It then reruns the clustering with stricter and looser thresholds and records how stable each cluster is. Isolates that join or leave with small changes are flagged as borderline. It drafts a summary for each cluster with its evidence. It never reports an outbreak. The scientist approves what is reported. Edge case: poor quality sequence is excluded from clustering and listed.

How it works

Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
Yes, continueYes, continueApprovedNoNo 1 STARTS WHEN New isolate sequences ready 2 USES A TOOL Check sequence quality and coverage 3 DOES Calculate distances to the existing collection 4 DOES Cluster at the standard threshold 5 USES A TOOL Compare each cluster with dates, places and sources 6 CHECKS THE RESULT Do dates and places fit a plausible link? If not: flag the cluster as genetic-only and ask forepidemiological data. Back to step 4. 7 USES A TOOL Rerun clustering with stricter and looser thresholds 8 CHECKS THE RESULT Is cluster membership stable across thresholds? If not: mark borderline isolates and test with anothermethod. Back to step 7. 9 DOES Write the cluster summary with evidence anduncertainty 10 YOU APPROVE Scientist approves what is reported 11 RESULT Cluster report
Read the steps as a list
  1. New isolate sequences ready
  2. Check sequence quality and coverage
  3. Calculate distances to the existing collection
  4. Cluster at the standard threshold
  5. Compare each cluster with dates, places and sources
  6. Do dates and places fit a plausible link?If not: flag the cluster as genetic-only and ask for epidemiological data. Back to step 4.
  7. Rerun clustering with stricter and looser thresholds
  8. Is cluster membership stable across thresholds?If not: mark borderline isolates and test with another method. Back to step 7.
  9. Write the cluster summary with evidence and uncertainty
  10. Scientist approves what is reportedThe agent waits here for your OK.
  11. Cluster report

How it decides

It calls a cluster stable when its membership stays the same across the threshold range. It marks isolates borderline when they move between clusters inside the range.

  • Exclude sequences below the quality cutoff and list them
  • Call a cluster stable if membership holds at plus or minus 2 SNPs
  • Flag a genetic-only link when no date or place connects the isolates
  • Use a second method for borderline isolates

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Distance threshold (default 3 SNPs)
  • Stability range (default plus or minus 2)
  • Quality cutoffs
  • Metadata fields used
  • Report format

What keeps you in control

It always asks you first

  • Any cluster report
  • Any notification to public health

Hard limits

  • Never report an outbreak or notify anyone
  • Always state the threshold used

It stops when

  • Done: clusters reviewed and approved
  • Stop: too few good sequences to cluster

Set it up

We guide you through the set-up, step by step

Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
  • The agent then walks you through connecting your own data, one source at a time
  • A downloadable copy with the flow chart, the rules and the full guide
Get access to this agent

An example run

What happensEight new Salmonella isolates were compared with 400. Five clustered within 3 SNPs of two earlier ones, from the same region over 3 weeks. At 5 SNPs, two more joined, so they were marked borderline. A second method kept one and dropped the other. One sequence failed quality. The scientist approved reporting a cluster of 8 isolates.

More agents for microbiologists