jean — Molecular Biology & Genomics

analyze_melt_curves

Read-only

Melt curves to melting temperature, under any of THREE engines that answer different questions and none of which is a fallback for another. plate (RDML-Python) normalises, derives twice, calls the peaks of -dF/dT and CLASSIFIES them against the target's expected Tm into per-reaction verdicts — one peak in the window is the product, an extra peak is an artifact or a primer dimer, no peak at all is a reaction that did not amplify; it is the only engine that produces a verdict. vertex (MBmca diffQ) fits a quadratic near the extremum and reports its vertex, so the Tm is NOT one of the sampled temperatures — the only sub-grid answer. curve (qpcR) does no plate normalisation and gives no verdicts, but reports every melting peak with the AREA under it. Replicate spread travels with all three, because it describes the readings rather than the method.

Input Schema

TableJSON Schema

Name	Required	Description	Default
`_files`	No	The project: record references shaped like _subject. Their records become library documents a tool names in sequences/parts/document; a trace or alignment is found by file name.
`engine`	No	which of the three engines runs. They answer DIFFERENT questions: `plate` is the only one that produces a verdict, `vertex` the only one whose Tm is not a sampled temperature, `curve` the only one that reports peak areas. Values: plate (Plate (RDML-Python) - normalise, classify, verdicts), vertex (Vertex (MBmca diffQ) - sub-grid Tm), curve (Per curve (qpcR) - peak areas).	plate
`_subject`	No	The record to act on: exactly one of sequence, accession (+source), sample (an id from list_samples) or file ({name, text} or {name, base64}); record_index picks one record of several.
`cut_area`	No	`curve` only: drop peaks whose area is under this fraction of the largest. 0 keeps every peak, which is upstream's own default. Only applies when engine is curve. Advanced.
`extremum`	No	`vertex` only: which extremum the melt sits at. `auto` reads each curve's direction, because upstream assumes a signal that RISES with temperature and a solution-phase melt FALLS — on which the default walks to the far end of the curve and returns a plausible-looking wrong Tm. Only applies when engine is vertex.	auto
`resample`	No	`vertex` only: resample the curve onto an even temperature grid before differentiating. Only applies when engine is vertex. Advanced.
`all_peaks`	No	`vertex` only: report every peak the curve carries, not the dominant one alone. Only applies when engine is vertex. Advanced.
`melt_table`	Yes	the melt table: a header row, first column the temperature and every other column one reaction's fluorescence. The same table the Melt tab takes.
`span_peaks`	No	`curve` only: the span Friedman's super smoother uses when the peaks are called. Only applies when engine is curve. Advanced.
`expected_tm`	No	the product's expected melting temperature, which is what turns a peak list into a verdict. Left unset, every peak is reported and none is classified - an honest absence rather than a guess. Read by the `plate` engine only, because it is the only one that classifies. In C. Only applies when engine is plate.
`norm_method`	No	`plate` only: how each curve is normalised before it is derived. Only applies when engine is plate. Advanced.	exponential
`peak_cutoff`	No	`plate` only: the height a candidate must reach, as a percentage of the largest peak. In %. Only applies when engine is plate. Advanced.
`span_smooth`	No	`curve` only: the smoothing span of the cubic spline fitted before the analytic derivative. Only applies when engine is curve. Advanced.
`fluor_source`	No	`plate` only: which matrix a peak's fluorescence drop is read from — the normalised curve or the smoothed raw one. Only applies when engine is plate. Advanced.	normalised
`expo_low_temp`	No	`plate` only: low end of the window the EXPONENTIAL background is fitted over. In C. Only applies when norm_method is exponential or combined. Advanced.
`peak_low_temp`	No	`plate` only: lowest temperature a peak is looked for at. In C. Only applies when engine is plate. Advanced.
`expo_high_temp`	No	`plate` only: high end of the exponential fit window. In C. Only applies when norm_method is exponential or combined. Advanced.
`peak_high_temp`	No	`plate` only: highest temperature a peak is looked for at. In C. Only applies when engine is plate. Advanced.
`peak_max_width`	No	`plate` only: widest a peak may be before it is not read as one product. In C. Only applies when engine is plate. Advanced.
`window_ceiling`	No	`vertex` only: the largest neighbourhood half-width the quadratic is tried at. Every width from 2 up to this is fitted and the one with the best adjusted R-squared wins, so the Tm arrives with the fit quality that chose it. Only applies when engine is vertex.
`true_peak_width`	No	`plate` only: half-width of the window around `expected_tm` a peak must fall in to count as the product. In C. Only applies when engine is plate. Advanced.
`replicate_robust`	No	replicate spread as median and MAD rather than mean and SD. Engine-independent: it describes the readings, so every run carries it. Advanced.
`replicate_percent`	No	report replicate spread as a PERCENTAGE of the mean rather than in degrees. Advanced.
`second_derivative`	No	`vertex` only: locate the inflection points of the melt (diffQ2) rather than the peak of the first derivative. Only applies when engine is vertex. Advanced.
`artifact_peak_width`	No	`plate` only: half-width used when an extra peak is measured as an artifact. In C. Only applies when engine is plate. Advanced.
`bilin_low_stop_temp`	No	`plate` only: end of the low seed window. In C. Only applies when norm_method is bilinear or combined. Advanced.
`bilin_high_stop_temp`	No	`plate` only: end of the high seed window — same narrow effect as its start, for the same reproduced reason. In C. Only applies when norm_method is bilinear or combined. Advanced.
`bilin_low_start_temp`	No	`plate` only: start of the LOW seed window the bilinear background is fitted from. In C. Only applies when norm_method is bilinear or combined. Advanced.
`bilin_high_start_temp`	No	`plate` only: start of the high seed window. NARROWER THAN IT LOOKS, and upstream's own quirk reproduced: the high baseline scan walks a hard-coded 90-91 C seed, so this knob feeds ONLY the first exponential-free fit and moves the result far less than the low window does. The windows actually chosen come back in the result. In C. Only applies when norm_method is bilinear or combined. Advanced.

Output Schema

TableJSON Schema

Name	Required	Description
`count`	Yes	how many reactions were analysed
`engine`	Yes	which engine produced these rows, named: the three answer different questions and a Tm without its engine is not comparable
`meanTm`	Yes	mean dominant Tm over the reactions that produced a peak
`noPeak`	Yes	reactions with no melting peak at all: a reaction that did not amplify, reported rather than dropped
`reactions`	Yes	one row per reaction, in the shape the CHOSEN engine reports: `plate` gives peaks plus a verdict, `vertex` gives the sub-grid Tm with the adjusted R-squared that chose its window, `curve` gives every peak with its area
`classified`	Yes	whether the peaks were classified, and why not when they were not — including 'this engine does not classify', which is a property of the choice and not a failure
`replicateSpread`	Yes	location and deviation per temperature per target, from the plate's own replicates. Engine-independent: it describes the readings, not the method, so every run carries it

Schema Changelog

Changes observed during successful MCP inspections.

First observedOct 20, 2026

TDQS

A4.1/5.0

Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare read-only/non-destructive analysis, and the description goes well beyond them: it discloses what each engine computes, that only 'plate' classifies, that 'curve' skips plate normalisation and produces no verdicts, and that replicate spread rides along with every engine because it describes readings not method. It stops short of describing performance or output shape, but the behavioural disclosures are substantial.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loads the key fact (three engines, different questions, no fallback) before enumerating each engine, so the agent gets orientation immediately. It is dense and delivered as one long paragraph, but every clause carries information that would otherwise be lost.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 29-parameter, three-engine tool with an output schema present (so return values need not be re-explained), the description supplies the engine semantics and the replicate-spread rule that an agent needs to choose correctly. The main omissions are cross-sibling positioning and any note on input-table requirements beyond what melt_table's own description states.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100% and each parameter already carries rich semantics (including 'Only applies when engine is X'). The description reinforces the engine-choice framing and the engine-independence of the replicate parameters, but this largely restates what the schema already documents, so the baseline of 3 holds.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb+resource ('Melt curves to melting temperature') and then precisely distinguishes the three engines by the question each answers, explicitly noting none is a fallback for another. An agent can tell the engines apart — and understand that this is a verdict-producing analysis vs. a raw peak profile — without opening any schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives strong intra-tool routing: 'plate' for verdicts, 'vertex' for a sub-grid Tm, 'curve' for peak areas, with the explicit warning that these are not interchangeable. It does not, however, contrast this tool with siblings like melt_profile or melting_temperature, so cross-tool selection is left to inference.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All jean — Molecular Biology & Genomics tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources

Need Help?