Service scopes sample numbers and data formats

A service quotation is checkable when it states how many samples will be accepted, what those samples are, which steps are included, which files come back, and which jobs are out of scope. Without those marks, "sequencing" or "proteomics" is a rumour about a study. This guide is how to write the marks. The broader commissioning decisions, including the scientific question and the reference, are in commissioning a sequencing or proteomics study. The enquiry pattern is how to write a laboratory sourcing enquiry. A scope can be discussed through laboratory procurement, through academic research for a university group, and by quotation request or contact. Discussing a scope is not the same thing as operating a sequencer or a mass spectrometer.

The decision the scope supports

You are deciding what you will later be able to point at when a file is missing or a step was skipped. Sample preparation in your own laboratory and an outsourced run are easy to blur. The scope draws the line. Everything on your side of the line is your SOP. Everything on the provider's side is a deliverable you can accept or send back. If you cannot draw the line, do not ship the samples.

Sample number and sample type

Sample number is an integer you can audit. Twelve tumours and twelve matched adjacent pieces are twenty-four samples if each piece is extracted separately. They are twelve samples if you pool them, and the scope should say you will not pool them. Replicates are samples. A control the provider asks you to include is a sample. Write the number, and write what you will do if intake QC fails one of them: replace it, proceed with the reduced number, or stop the batch. Silence on a failed sample is how a sample disappears from a table without a sentence.

Sample type is the material as it will be handed over. Fresh frozen tissue, cultured cells, already extracted RNA, formalin-fixed material, a protein digest, or a serum aliquot are different types. They imply different included steps. Name the species and the matrix. Name the tube and the labelling rule. Name the temperature class at handoff, using the storage logic in your own SOP. A provider cannot invent the biological identity of a mislabelled tube. Your label is part of the scope.

Also state the minimum quality you will meet or the minimum the provider requires, in their words, once they have written it. A promise that every sample will be "good" is not a quality line. Concentration, integrity, or a volume, as the method needs, can be. Follow the provider's written intake rule and your SOP. Public method notes such as protocols.io show how much extraction choices change what "RNA" means. They are not your scope. Your scope says which of those choices you already made.

Included steps, marked one by one

Write four questions even if the answer to some is no.

Extraction: will the provider extract nucleic acid or protein from the material you send, or are you sending material that is already extracted?

Library or digest preparation: will they build the sequencing library, or perform the protein digestion and cleanup, or are you sending a library or a peptide mixture?

Measurement: will they sequence, or acquire mass spectra? Say the read type or the acquisition style at the level you understand and they can confirm, without pretending a brand of instrument is the science. The scientific need might be short reads for expression, or tandem mass spectra for identification. Let them confirm the configuration in writing.

Analysis: will they align, quantify, or search a database, or will they return raw files only? Analysis is a separate deliverable. Leaving it implied is how a folder of reads arrives and everyone looks at each other.

Each yes needs a short definition. "Library included" should say DNA or RNA, and whether strandedness or a selection method is part of the offer, if your design depends on it. Each no should be an explicit out-of-scope line so it cannot drift back in.

Data format

Name files a person can look for on a disk. For sequencing, FASTQ is a common raw read format. An alignment file and a count matrix are processed formats. For proteomics, vendor raw spectra and an identification table are different objects. A PDF report is a narrative, not a substitute for the raw file if you asked for the raw file.

Say how the files will be transferred and for how long the provider's copy remains available, if they have stated a practice. Retention details belong in the statement of work beside ownership. Here, at least name the format so "we sent the data" can be tested. If you cannot open the format, it is the wrong format to accept, and you should say so before the run.

A worked RNA expression scope

You have cultured cells, two treatments, four biological replicates each. That is eight samples, not two. You will extract RNA yourself and check integrity in your laboratory. The scope therefore marks extraction as out of scope. You will send eight tubes of RNA, labelled with the treatment and the replicate, on the frozen class your SOP uses for RNA.

Library preparation: included, stranded RNA library, from the RNA you send, with the provider's written minimum input. Sequencing: included, short reads, with the read length and depth the provider writes in the quotation after you ask for those numbers. You do not invent a depth here and call it universal. You require the quotation to state the figure it will deliver. Analysis: not in this scope. You want FASTQ files and a simple sample sheet that matches your labels. Alignment and statistics will be a later line if you commission them. Out of scope: growing the cells, deciding biological meaning, and any clinical interpretation.

The quotation comes back saying "RNA-seq, eight samples, data delivered". You mark library, sequencing configuration, and FASTQ as silent. You ask for a revision that says yes to library preparation, states the read length and depth they will deliver, and names FASTQ. You do not ship on the short sentence. If one RNA fails their intake, the scope already says you will replace that replicate rather than drop it quietly, because four versus three replicates changes the design you registered with yourself.

Protein identification would be the same skeleton with different steps: digestion included or not, spectra as the raw format, a search database named if analysis is included, and UniProt as one public example of a protein sequence resource the search might use. Name the database version in the scope if they will search. Do not assume a default.

The branch when the count changes

If you add samples after the quotation, the scope has changed. Ask for the number, the type, and the formats to be rewritten. A verbal "send two more, it will be fine" is how the extra tubes miss the library batch.

If intake fails and your rule was to stop, stop. Do not let a reduced data set inherit the original conclusion. If your rule was to replace, send the replacement under the same labels' logic and record the swap. The sample sheet is a controlled document, not a spreadsheet nobody owns.

Failure modes

The first is the bundled verb. "Full RNA-seq" sounds like extraction, library, sequencing, and statistics. Unless each is marked, you have a slogan. Break the verb before you send tissue.

The second is the format you cannot open. A proprietary file without a raw export locks the reanalysis you thought you had reserved. Name the open format in the scope.

The third is scope creep in the provider's favour or in yours. They skip a library QC you thought was included, or you send twenty samples against a quotation for eight. Both are the same failure: the written number was ignored. Stop and revise.

Research use and biosafety

State that the work is research, and state the biological risk as your institution classifies it. The provider can agree to handle material under conditions you specify. They do not approve your biosafety level, and you do not approve theirs by posting a parcel. Follow both institutions' rules. A data file is not a diagnostic report. Do not write clinical claims into a research scope to make the format list sound important.

Samples waiting on the bench

Heat and a power cut threaten the samples before they ever meet a provider. The scope should name the storage class you will hold and the condition you will ship, and your building has to be able to hold that class. A scope that assumes a freezer nobody is monitoring is a sample-type fiction. Write the handoff you can actually perform. Do not add a guessed turnaround. If timing matters, ask for it to be written on the quotation, and treat silence as unknown.

What to send

Send the integer, the sample type, each step marked yes or no, the file types, the failed-sample rule, and the out-of-scope list. Ask the quotation to repeat those marks. Point the scientific question at the commissioning note, and point the commercial file at laboratory procurement or academic research. Reach a person through contact or a quotation request. A scope can be discussed. Ship only after the repeated marks match the design you are willing to interpret.