Applied AI-Powered Bioinformatics Tools

Energy Expenditure Study Summarizer

Last reviewed

Beginner

What you'll learn

~40 min
  • Ask an AI tool, in plan mode, for a program that reads a doubly labelled water study's results table (one row per participant, as your laboratory's own calculation writes it) and your laboratory's limits sheet
  • Read its plan for what matters: the two ratios (dilution spaces, elimination rates) worked out at full precision and compared with your written limits, every row's note shown whether it has a flag or not, a participant code that appears twice shown twice, and no row called passed
  • Check its list against values worked out for the practice files
  • Say why a row just outside a limit and a row just inside it need the same look, why a row with no flag can still need one, and who decides what happens to a participant's result

The story, and what you will end up with

It is Tuesday 10 November 2026. You work in a university stable-isotope laboratory that measures energy expenditure by doubly labelled water for other research groups. One study, 18 adult participants, has reached its last step in the lab. The urine samples have been analysed on the mass spectrometer, the duplicate analyses and reference waters checked, the dose records checked, and the laboratory’s own calculation has been run. That calculation, which the lab has checked and keeps under its procedures, works out each participant’s elimination rates and dilution spaces from the isotope measurements, carbon dioxide production from those, and energy expenditure from that. A blank sample time, or one out of order, never reaches its results: the lab’s own intake checks catch those before the calculation runs.

What it writes is a results table, one row per participant. Before that table goes back to the study’s investigators, someone in the lab reads it row by row, notes included, against the limits the lab has written down. Until now you did that by eye, in a spreadsheet.

By the end of this lesson you will have a small program on your own computer, the results check, that reads the study’s results table and the lab’s limits sheet and makes a page you can print: which participants’ rows someone needs to look at, and why, with the value and the limit beside each, and every row’s note beside it, flagged or not. It calculates no elimination rate, no dilution space and no energy expenditure, and it decides nothing about any participant. You build it by asking your AI tool in plain English, and you check it against figures this lesson gives you.

🔍Words used in this lessonclick to expand
  • Doubly labelled water: water carrying extra deuterium (heavy hydrogen) and extra oxygen-18. A participant drinks a measured dose, and urine samples over the following two weeks show how fast each isotope leaves the body.
  • Elimination rate, kD and kO: how fast deuterium (kD) and oxygen-18 (kO) leave the body, per day. In the introduction of Speakman and colleagues’ paper (Cell Reports Medicine, 2021, read 11 November 2026), labelled oxygen “is eliminated as both water and CO2”, deuterium “will be predominantly eliminated as water”, and “The difference in elimination rates of the two isotopes (hence “doubly labeled” water) gives a measure of rCO2” (carbon dioxide production). The elimination rate ratio is kO divided by kD.
  • Dilution space, ND and NO: the amount of body water each isotope mixed into, in kilograms. The dilution space ratio is ND divided by NO.
  • Enrichment above the pre-dose sample: how much more oxygen-18 a sample carries than the participant’s own sample from before the dose, in per mil (‰, parts per thousand on the scale isotope labs report in).
  • Limits sheet: the lab’s written list of the range each check must fall in. It is the lab’s own; this lesson’s is a practice one.
  • Plausibility range: the lowest and highest energy expenditure this practice lab accepts without a second look. It is the practice lab’s own, described in the box below.
  • Flag: a mark on a row saying a person needs to look at it, and why. It is not a decision.

Before you start

First, if your AI tool is still open, leave it: type /exit and press Enter (if you use Antigravity, press Ctrl+D with nothing typed instead), and wait until you see the $ or % prompt. The next line goes there, in the terminal, not in the AI tool. It makes a folder for this lesson beside your practice folder and goes into it:

Terminal window
mkdir -p ~/bio-workbench/12-dlw-review && cd ~/bio-workbench/12-dlw-review

Start your AI tool with the start line you kept from the setup lesson, the one with approval prompts off. Then ask it to set the folder up:

Set this folder up so we can undo any change you make, download https://learn.intellimetrics.net/starters/biotech/12/dlw-results.zip and unpack it into a data folder, and save a checkpoint.

Look for it saying three things: the folder is set up with Git (the program that saves snapshots, from setup), two files are in the data folder (participant-results.csv and lab-limits.csv), and a first snapshot, your checkpoint, is saved. Later you can ask whether the files have changed since then. If it does not mention one of them, ask: “Did you save a checkpoint with the files in it?” It may also keep the zip file in the data folder, which is fine; if it says it kept one somewhere outside this folder, ask it to delete that copy. If the download does not work, download dlw-results.zip in your browser, check where it went (the browser’s list of downloads shows the folder), and tell your AI tool, for example: “dlw-results.zip is in my Downloads folder. Unpack that one file into a data folder here, and save a checkpoint.” In Windows PowerShell or a desktop app, “my Downloads folder” works as is; only say “my Windows Downloads folder” if you use Ubuntu (WSL) on Windows, which keeps its own Downloads folder apart from Windows.

The box below says what is in the practice files; later steps refer to them.

ℹWhat is in dlw-results.zipclick to expand

Two files, each of which opens in a spreadsheet program, from one made-up study:

participant-results.csv, the results table as the practice lab’s calculation wrote it: 19 rows for 18 participants, P01 to P18. Each row has the participant code, body mass, the date and time of the pre-dose sample, the dose and four samples after it (planned for days 1, 5, 10 and 14 after the dose), kD and kO, ND and NO with their ratio to three decimal places, the oxygen-18 enrichment of the last sample above the pre-dose sample, the energy expenditure in kcal a day, and a column for the lab’s notes.

lab-limits.csv, the practice lab’s written limits:

CheckLowestHighest
Dilution space ratio, ND divided by NO1.0001.070
Elimination rate ratio, kO divided by kD1.11.7
Oxygen-18 enrichment of the last sample above the pre-dose sample8 ‰none
Energy expenditure, this practice lab’s own plausibility range1,000 kcal a day6,000 kcal a day

The practice lab took the first three from the International Atomic Energy Agency’s guide (quoted under “How real laboratories differ”). The fourth row is the practice lab’s own, taken from no source; it is neither a physiological range nor a standard. The lab’s rules on its limits: a value exactly on a ratio limit or on the energy range counts as inside, and, because the guide says the enrichment should “exceed” 8 ‰, an enrichment of exactly 8 is a flag. No practice value sits on a limit, so the files cannot show how your program treats one; a step further down checks the enrichment case with a made-up row. The lab’s one other rule is its own too: each participant code appears once.

The rows you will meet further down:

ParticipantWhat is in the row
P03ND 40.22 kg, NO 37.31 kg, ratio 1.078
P05Energy expenditure 13,226 kcal a day; body mass 66.1 kg; both ratios and the enrichment inside their limits
P07Two rows. The second, last in the file, has the note “re-analysed 2026-10-30” and slightly different kD, kO and energy expenditure (2,706 kcal a day in the first row, 2,677 in the second)
P09kD 0.1180, kO 0.1270 a day; energy expenditure 496 kcal a day; body mass 58.0 kg
P11Every figure inside its limits; something written in its notes column
P12Enrichment of the last sample 6.9 ‰
P13ND 41.16 kg, NO 38.40 kg, ratio 1.072

Everything in these files is made up for this course: the study, the participants and every value; no row describes a real person or a real study. Each lesson in this track has its own practice files, with their own people and dates.

Switch to plan mode

Your AI tool is already open. Switch to plan mode now: Shift+Tab until it says plan, or /plan (the keys for each tool). In plan mode it usually reads your files, asks its questions and shows you a plan before it builds anything; the order can vary.

The conversation

Each message below is an example. Say it your own way.

Say what you want

Say what you do, which files to use, and what you want from them:

I check doubly labelled water results in our lab before they go back to the study team. Using the files in data, show which rows need a look and why, and every row's notes. Ask me what you need first.

Your AI tool usually reads both files and comes back with questions. It may also point out some of the rows this lesson is about before you ask.

Answer its questions

Answer each in a sentence. These are questions the tools often ask, with one way to answer each. Yours may ask them in a different order or show a list of choices:

If it asksYou might say
Only the limits file, or other checks too, such as sample order or duplicates?Every limit in lab-limits.csv, and any code that appears twice.
Any other review rules?No other rules.
P07 appears twice. Use the re-analysed row?Neither. Show both rows and choose nothing; I’ll check the analysis log.
P11’s note says part of the dose was spilled. Should a note count as needing a look?Not as a flag. Show every note beside its row; I read them.
What do you mean by every row’s notes?The lab’s own note from the file, beside each row.
A table on screen, a saved file, or a page?One page I can open in my browser and print. A program I can run again.

If it asks several questions at once, answer the ones about what is checked and against which limit from your lab’s written procedures, a sentence each, and accept its suggestions only about how the page looks. If you cannot tell which kind a question is, treat it as a rule. If it asks something not listed here, answer it the same way, in a sentence.

Give it the lab’s two rules

The files do not hold two of the practice lab’s rules, so say them in a message of their own:

Two rules: work both ratios out from the two columns at full precision, and call a row without a flag "no flag", never passed, approved or valid.

Wait for its reply, then read the plan.

Read the plan

Next it usually shows you a plan: what it will build and how it will check it. Read it for six things:

  1. Every row is checked against every limit in lab-limits.csv, with the limits read from that file, so your lab can change them without changing the program.
  2. The two ratios are worked out from the columns at full precision, that is from the two spaces and the two rates, not from a rounded figure, and nothing else is calculated: energy expenditure is compared with the sheet’s range as your lab’s calculation gave it.
  3. Every row’s note is shown beside it, whether the row has a flag or not.
  4. A participant code that appears twice is shown with both rows, and neither is chosen.
  5. Each flag says which check, the value and the limit, and a row with no flag is called “no flag”, never passed, approved or valid.
  6. It is a program you can run again on the next study’s files, it sends nothing, and your two files are read and left as they are.

A first plan may still leave one of these out, most often item 2: a plan may compare the ratio column in the file, which is rounded, instead of working the ratio out. Say so in a sentence, in your own words, for example:

Work the space ratio out from ND and NO, not from the ND_NO column.

If item 6 is missing, say: “Make it a program I can run again on the next study’s files.”

When the plan has all six, approve it: if your AI tool offers a choice to go ahead, pick it; otherwise say “Yes, go ahead and build it.” In Codex, press Shift+Tab first to leave plan mode. In Copilot, saying it brings up a choice to go ahead; pick that. Let it build. It may run what it built and tell you how to run it yourself; either way, the next step asks it to. When it has finished, ask it: “Did you install anything, and where?” Nothing, or something inside this lesson’s folder, is fine for these lessons. If it names anywhere else on your computer, such as the zip file it downloaded, ask it what that is for, and to remove it if you do not want it there.

Look at what it built

Ask it to run what it built:

Run it and show me which rows need a look.

It may show the list in the terminal and say where the page is saved, or start the page and give you an address to open. To see the page itself, ask: “Open the results check in my browser.” If it gives you a file location or an address without opening it, copy that into your browser’s address bar and press Enter. If it starts the page at a local address, one ending in a number, and says the address is already in use, say: “Something else is using that address. Don’t stop it. Use a different address for my page.” If your AI tool does not take a message after opening the page, wait a minute. If it still does not, close the terminal window, open a new one, type this lesson’s folder line again, start your tool, and say: “Start the page again so it keeps running while we talk.” The page and your saved work are files in the folder, so nothing is lost. When you have finished with the page, ask: “Stop the results check page if it is still running.”

Your page will be worded and laid out in its own way. Check it for these:

  • It reads 19 rows for 18 participants. Six participants have flags: P03, P05, P07, P09, P12 and P13. The other twelve have no flag, though one of them, P11, still needs a look for its note.
  • P03’s dilution space ratio is 1.078 and P13’s is 1.072, both above 1.070. If P13 has no flag, the ratio may have been rounded to two decimal places (1.07) before it was compared; tell your AI tool to compare at full precision.
  • P09’s elimination rate ratio is 1.076 (0.1270 divided by 0.1180), below 1.1, and its energy expenditure, 496 kcal a day, is under the plausibility range.
  • P05’s energy expenditure, 13,226 kcal a day, is over the plausibility range, though its two ratios and its enrichment are inside their limits.
  • P12’s enrichment of the last sample is 6.9 ‰, not above 8.
  • P07 is shown with both rows, the second with its “re-analysed” note, and neither is chosen.
  • P11 has no flag, and its note is shown beside it.
  • No row is called passed, approved or valid.
  • When you ask “How does it treat a value exactly on a limit? Show me with one made-up row in a copy, its enrichment exactly 8, then delete the copy.”, it says how and shows you the made-up row. The practice lab flags an enrichment of exactly 8. If the made-up row has no flag, the program counts 8 as inside; say: “The enrichment must be above 8. Flag exactly 8, and show me the made-up row again.” If a value exactly on another limit matters in your lab, ask it to show you the same for that limit.
  • When you ask “Does the program name any participant from these practice files? If it does, make it work those out from the files instead.”, it says it does not, or changes the program so it finds them from the files and tells you what it changed. A program built in a conversation may keep this study’s answers written into it, and on the next study it would flag the wrong rows.
  • When you ask “Have my results files changed since the checkpoint?”, it compares them with the checkpoint and says no.

If something on the list is not true, tell your AI tool in a sentence what you see and what you expected, for example: “P05 has no flag. Its energy expenditure is over the top of the plausibility range in the limits sheet, so flag it.”

Notice what a limit can and cannot tell you

P13’s ratio is 1.072, about two thousandths over the limit. Try making one of its spaces smaller by about half of one per cent:

Make a copy of the results file in this folder. In the copy only, change P13's ND to 40.96 and its ratio to 1.067. Then show me P13's flags from the copy beside the original.

Before you read on: P13’s deuterium space is now 0.2 kg smaller. What do you expect the list to say?

P13 has no flag in the copy. Nothing else about the row changed. 0.2 kg is about half of one per cent of a 41 kg space. For comparison, the International Atomic Energy Agency’s guide (section 6.4, on a laboratory’s own precision study, using “actual specimens for which sufficient volumes are available for five analyses” or mock ones) says a laboratory “should be able to determine TBW with a relative standard deviation of 0.5–1.5%” (TBW is total body water). So the original row and the copy differ by a change in one space about the size of the precision the guide expects of a laboratory’s body water figures. The flag on the original P13 says where to look; the copy’s missing flag is not an approval.

Now look at P09, flagged twice: its elimination rate ratio is under 1.1, and its energy expenditure, 496 kcal a day for an adult of 58 kg, is under the plausibility range. Its two rates are close: kO 0.1270 and kD 0.1180 a day, a difference of 0.0090. Make kO one per cent smaller, 0.12573, and the difference falls to 0.00773, about 14 per cent smaller. A difference between two close numbers moves by a large share when either moves a little. P05 is flagged once, for 13,226 kcal a day, though its rates and spaces look like any other row’s. That looks like a figure that went wrong after the calculation, such as a result written in kilojoules or a slip in copying; what happened is for the laboratory manager to find out from the calculation’s own output. Each flag sent you to a row; reading the row is what tells you what kind of problem it is.

Then find P11 on your page. It has no flag. Read its note: “part of dose spilled at dosing; amount not recorded”. The calculation worked from the full dose in the dose record, and the guide (section 4.2) ties a dilution space to “knowledge of the exact dose of deuterium labelled water consumed”. How far the spill moves P11’s figures is a question for the laboratory’s calculation, not for this list. Every figure in P11’s row is inside its limits, so no limit in the sheet flags it; only someone reading the note can see the spill.

So that no row with a note slips past, give notes a mark of their own:

Mark every row that has a note "note to read", apart from the flags, and show me which rows that is.

It should mark two rows, P11 and P07’s second row, and leave the flags as they were: P11 still has no flag.

Re-analysing P09, going back to the analysis log for P07, checking P05’s figure against the calculation’s own output, or recalculating P12 without its last sample, which shortens the period its rates are worked out over: each is a decision the laboratory makes and records in its own procedures. What P11’s note means for its result, the laboratory works out from its dosing record and settles with the study team.

A limit shows some of the rows to look at, not all of them and not which results stand: re-analysis is the laboratory manager’s decision, and whether a result is reported or excluded is settled with the study’s investigators under the study’s own rules, never by a list of flags.

When you have looked, remove the copy: “Delete that copy and anything you made from it; I’m done with them.”

One real run

Here is part of what one AI tool said when it was asked to run what it built, with Claude Code on 11 November 2026. Yours will word it differently; check your page against the list above, not against this.

The other 12 rows show "No flag". P11 is one of them, and its note, "part
of dose spilled at dosing; amount not recorded", is shown beside it on the
page.

The participant and the note it talks about are made up for this course.

How real laboratories differ

The three limits in the practice sheet come from the International Atomic Energy Agency’s guide, IAEA Human Health Series No. 3, Assessment of Body Composition and Total Energy Expenditure in Humans Using Stable Isotope Techniques (2009, read 11 November 2026), section 6.4.2, its checks for studies with several samples after the dose. Its words: “The ratio of the deuterium space to the 18O space should be between 1.000 and 1.070 with the ideal ratio for adults being 1.034. If the ratio is outside this limit, then the results are suspect.”; “The sixth quality control step is to consider the ratio of the isotope elimination rates (kO:kD). This should be between 1.1 and 1.7.”; and “The enrichment of 18O should exceed 8‰ or 16 ppm excess.” (18O is oxygen-18; ppm excess is parts per million above the pre-dose sample.) Where a result is suspect, it says “the data input should be checked and if necessary the isotopic analyses should be repeated.”

  • Your limits are your lab’s. These are the guide’s checks for its protocol with several samples; the 1.034 it calls ideal is for adults, and its section on special populations (6.5) shortens the sampling period for infants and preschool children. In its checks for the two-point protocol (6.4.1: a pre-dose sample, one on the day of dosing and one at the end of the study) it ties the 8 ‰ to the instrument: the final enrichments “should exceed 60 times the measurement precision or about 8‰ or 16 ppm excess for a highest quality laboratory.” For the space ratio in very small children, the paper below reports in its summary that it “is lower at low body masses”, which it gives as under 10 kg. The same paper says, in its methods: “There is no standard approved protocol for the use of the DLW technique”. So the limits sheet the finished program reads is the one in your lab’s written procedures, at its current version.
  • More checks come first, in the lab’s own calculation. The same section of the guide lists others in its own words: “check the internal precision of the duplicate or triplicate analyses of each isotopic abundance analysis”; “check R2 of the regression line through the enrichment data”, and if it falls short, “check the date and time of the samples and for errors in data input.”; and “compare the isotope elimination rates calculated by dropping one data point at a time”. For deuterium it gives a figure for one way of preparing samples only, equilibration: “If the equilibration technique is used to prepare hydrogen samples for analysis, the final enrichment should be greater than 128‰ or 20 ppm excess.” Its next sentences, after that deuterium figure, say “Specimens with enrichments below this level should not be included in the calculation of kX and ΔBW” (the elimination rates and the enrichment measured in body water). Sending P12’s low oxygen-18 figure back to the lab’s calculation, not to the study team, is this practice lab’s own rule. If your lab’s calculation already marks rows outside the limits in your sheet, the results check is a second look at the table as it will go back, not a replacement for those marks.
  • Which equation the calculation uses is the lab’s choice too. The practice lab’s calculation uses Equation 1 of Speakman and colleagues, “A standard calculation methodology for human doubly labeled water studies” (Cell Reports Medicine, 2021, read 11 November 2026), whose authors write, at the end of their section on children, adolescents and adults, “we recommend adoption of Equation 1 in future studies of children, adolescents, and adults to derive rCO2” (carbon dioxide production), and, opening their next section, on small infants and babies, “The recommendation above refers to subjects aged ≥2 years.” The finished program does not use or check any equation.
  • A real results table may hold more. Check what yours holds, for example: sex and age, the dose given and its enrichment, the food quotient the calculation used to turn carbon dioxide production into energy expenditure, several analysis runs, and the participant’s own record of the study fortnight: a dose spilled, a void missed or timed from memory, an illness, travel, a change in what they drank. Check whether your procedures have that record read beside the row, the way you read P11’s note; the practice files hold only that short notes column.
  • Sample times carry more than order. A sample can be in order and still far from its planned day, and a time a participant wrote down may be approximate. Ask your analyst how your calculation handles such a time, and check what your procedures say.
  • A plausibility range, if your lab uses one, comes from your procedures. If they compare each result with what is plausible for the participant, that range goes in your limits sheet.
  • A code that appears twice can mean the export took in more than one run. Your analysis log says which run is current; the list only shows that there are two.
  • What goes back, and with which comments, is in your procedures. Nothing in this lesson drops a flagged row from what the study team receives or calls a row with no flag valid; what the lab sends, and how, follows the study’s approvals and the lab’s procedures.

Keep going

Small changes do not need plan mode. Say each in a sentence, then run it again:

  1. “Also flag any sample taken more than a day away from its planned day: days 1, 5, 10 and 14 after the dose.” On the practice files this flags nothing: every sample is within a day of its planned day.
  2. “Show how far each flagged value is from its limit, so I can see which rows are only just over.”
  3. “Add a place beside each flagged row for me to write what was decided, by whom, and when.”
  4. “Save a checkpoint.” Your AI tool saves a snapshot you can return to.

Your own study

When you want to check a real study’s results, use the finished results check. It runs on your real files by itself, with your AI tool closed, so your AI tool never sees them; never start your AI tool in their folder. Nothing confidential or unpublished, and nothing about people or about animals under a protocol, goes to your AI tool or into this course folder: not the file, not a row, not a value typed into a message. Your lab’s calculation will not write a file shaped like the practice one: its columns will have other names, and it may carry body mass, dates and notes that, in a small study, could point to a person. Your limits may not be a file yet; you type them into a blank sheet. Do these steps in order. Steps 1 to 4 happen now, while your AI tool is open and this folder still holds only the practice files:

  1. Go through the data check, the questions in setup about what the data is, who decides how it is used, and which AI tool would see it. Participant results belong to the study, under its approvals; your laboratory manager and the study’s investigators decide how they are handled. Go on only if they allow this use.

  2. Decide on a folder of its own for your real files, outside this course folder and somewhere the study’s approvals allow its data to be kept, for example ~/dlw-review. Nothing has to be in it yet. If the folder is not there, ask your AI tool to make it: “Make an empty folder called dlw-review in my home folder.”

  3. Ask your AI tool for a blank limits sheet, for example:

    Make a blank limits sheet my spreadsheet program can open, in ~/dlw-review. Show the check working with a made-up limit for each check and one made-up participant row, then clear the rows.

    Use your own folder’s name, written out in full from your home folder. Look for where it says it saved the sheet, check that it says the made-up row was flagged or not as the made-up limits would have it, and that the sheet is back to its headings only. It may say that the program recognises each check by the first words of its name, such as “Dilution space ratio”. If it does, note those words, and start each check’s name with them when you fill the sheet in. If it offers to copy the practice limits into it, say no: your limits come from your lab’s procedures.

  4. Some rules are written into the program, not the sheet: how a value exactly on a limit counts, that each participant code appears once, which rows get a “note to read” mark. Where your lab’s written procedures differ from the practice lab’s, tell your AI tool each difference in a sentence. Wait for its reply, then ask: “Show me one made-up row checked under my rules, then delete anything you made for it.” Check that what it says agrees with your procedures, and say “Save a checkpoint.”

  5. Leave your AI tool the way you did at the start of this lesson. Open an export from your lab’s calculation in your spreadsheet program, note its column names and how it writes dates and times, and close the export. Then type this lesson’s folder line again and start your AI tool.

  6. Tell your AI tool those column names, and give it two rows you make up in the same shape, with made-up codes and values, never a real row. For example: “My results export has the columns Subject, Dose Time, Urine 1 to Urine 6, kd, ko, Nd, No, Nd/No, Final 18O, TEE and Comment, with times like 09/14/2026 08:05. Here are two made-up rows in the same shape.” Then paste the two rows you made up. Wait for its reply, then say “Change the results check to read files like that.” Wait for its reply, then say “Save a checkpoint.”

  7. Ask your AI tool for the exact line, for example: “Write me the exact line to run this myself on the files in ~/dlw-review, saving the page there.” Use your own folder’s name, written out in full from your home folder as above. If it offers to copy your real files into this folder instead, say no. Copy the line it gives you into a note. It will look something like this one, though yours may differ:

    Terminal window
    python3 /home/dana/bio-workbench/12-dlw-review/scripts/check_dlw.py /home/dana/dlw-review/participant-results.csv /home/dana/dlw-review/lab-limits.csv /home/dana/dlw-review/dlw-check.html

    Read it left to right. python3 runs a Python program: first comes the results check, in this lesson’s folder; then the study’s results export and your limits sheet, both in your folder for real files, dlw-review; last, the name of the page it writes there. /home/dana is the home folder of someone whose user name is dana; your own user name goes where dana is, and on a Mac the folders begin /Users/ instead of /home/. For each study you change the file names to that study’s. The line runs the program where it is, so leave this lesson’s folder in place. If you are unsure about any part, ask your AI tool now, while it is still open.

  8. Leave your AI tool the way you did at the start of this lesson.

  9. Fill in the limits sheet from your lab’s written procedures. Before you use the results check on a new study, run it on one you have already reviewed by hand: save that study’s results export into your folder with the name the line uses, paste the line from your note at the $ or % prompt and press Enter. The page appears beside your real files, not in the course folder. Compare its flags and marks with your own review, row by row, the same way you checked the practice files. To come back to the lesson folder later, type this lesson’s folder line again, the line near the top of this page that begins with mkdir.

  10. Before its flags go into any record, your lab’s procedures decide how a new checking tool is approved and kept.

Try another case

Before the quiz, write two or three sentences for yourself. The results check has flagged P09 and P12 in a real study. Before the table goes back, one of the study’s investigators asks you to send only the rows with no flag, so the team can start its analysis. What do you do, and what do you tell them?

A good answer does not drop rows because they carry a flag, and does not send the table as if the rows without one were cleared. It looks into P09 and P12 first in the lab’s own records (the participant’s record of the fortnight, the duplicate analyses, the calculation’s fit) and decides, as the lab, whether to re-analyse. It reads the notes on the rows with no flag too. It tells the investigator how the lab’s procedures send results back, with the lab’s comments, and that leaving a participant out is decided with the study team under the study’s own rules, not by the list.

Key takeaways

  • A short ask is enough. Say what you check, which files, and what you want; its questions fill in your lab’s rules.
  • Your limits come from your lab’s written procedures, never from the AI tool’s suggestions, and the program reads them from a sheet you keep.
  • The program works out no rate, space or energy expenditure. It works the two ratios out again from the columns, compares them and the reported figures with your limits, and checks codes.
  • A flag marks some of the rows that need a look, not all of them. A row just over a limit and a row just under it can differ by about the precision expected of the measurement, and a row with no flag can carry a note that needs reading, as P11’s does: a flag says where to look, and a missing flag approves nothing.
  • Practice data stays practice data: real participant results go to the finished results check, with your AI tool closed, after the data check.
KNOWLEDGE CHECK

In a study, one participant's dilution space ratio is 1.072 and the results check flags it. Another's is 1.068 and has no flag. Your lab's limit is 1.070. What does the list tell you about the two results?

What’s next

The next lesson is about a training mock case’s custody records: each item’s hand-overs laid out in order, with what needs a look marked.

intellimetrics · Arlington, VA
© 2026 intellimetrics.net
Search lessons