University of Turku Food Development Master’s Thesis Guide: DEKE0122, 30 ECTS, Experimental Food Research and UTUGradu

Quick answer: what is the University of Turku Food Development thesis route?

The current University of Turku Master’s Degree Programme in Food Development is a 120 ECTS, two-year Master of Science (Technology) programme in the Faculty of Technology. The controlling Peppi programme is FODEDI2427 / 96204. Its curriculum contains a 40 ECTS “Master’s Thesis in Technology and supporting studies” package, but the actual thesis is DEKE0122 Master’s Thesis in Technology, Food Development, 30 ECTS. The package also contains DEKE0120 Introduction to Master’s Thesis in Technology, 5 ECTS, DEKE0121 Master’s Thesis Seminar, 5 ECTS, and TTDK1308 Degree Qualifying Examination, 0 ECTS. This distinction matters: do not call the thesis itself 40 ECTS.

1. Read the 120 ECTS degree correctly

The degree contains 10 ECTS of Common Studies, 30 ECTS of Food Development studies, the 40 ECTS thesis-and-supporting-studies package, a 20–25 ECTS thematic module or minor, and 15–20 ECTS of other studies including mandatory language studies. The ranges are planning flexibility, not permission to redefine the degree. Use the current personal study plan to make the total exactly 120 ECTS.

2. The actual thesis is DEKE0122, 30 ECTS

DEKE0122 is the current Food Development Master’s Thesis in Technology. It is an Advanced Studies course, taught in English and graded 0–5. The programme page describes the final project as independent experimental laboratory research under supervision. The thesis is written from the experimental results and the relevant literature.

3. The 40 ECTS package is four separate components

Peppi decomposes the package into 5 + 5 + 30 + 0 ECTS. DEKE0120 is the planning course, DEKE0121 is the thesis seminar, DEKE0122 is the thesis itself, and TTDK1308 is the maturity/degree-qualifying examination. Keeping these components separate prevents the most common structural error in this programme.

4. DEKE0120 is the pre-experimental planning gate

DEKE0120 must be taken before the experimental thesis work begins. The student reads thesis-related literature, writes an essay, and prepares an independent 2–4 page research plan after discussion with supervisors. That plan should state the background, aims, methods, schedule, expected results and implementation risks. Treat it as the operational contract for the experimental phase.

5. Make the research plan executable, not ceremonial

Convert every objective into a measurable output. Identify samples, experimental units, controls, instruments, methods, primary response variables, statistical comparisons and fallback options. A plan that says “analyse food quality” is too vague. A stronger plan says which quality dimensions are measured, with what method, at which time points, and how the resulting values answer the research question.

6. The DEKE0122 experimental phase is described as four months

The current course says the 30 ECTS thesis begins with a four-month period of experiments and measurements under an accepted research plan. That is the experimental/measurement phase, not a promise that all planning, analysis, writing and seminar obligations fit into four months. Build the calendar backward from instrument access, sample availability, biological growth or fermentation time, and the seminar schedule.

7. Supervision does not transfer responsibility away from the student

DEKE0122 expects close cooperation with a supervisor and possibly a research group, but the student remains responsible for implementing the research plan, reporting progress and proposing changes when results require them. Keep a decision log for method changes, exclusions, failed batches and deviations from the original plan.

8. Use the required scientific-report architecture

The current course description explicitly identifies Introduction, Materials and Methods, Results and Discussion. The Introduction establishes prior research and the aims. Materials and Methods should make the work reproducible. Results should report findings without hiding inconvenient outcomes. Discussion should compare the findings with literature and explain their wider significance without overclaiming.

9. The current Food Development page expectation is typically 50–60 pages

DEKE0122 says the thesis typically contains 50–60 pages and follows a defined format with separate detailed formatting instructions. Treat that as a current programme-course expectation, not a universal University of Turku rule. If department instructions change, the current instructions win over an older template or a previous student’s PDF.

10. DEKE0121 is a real separate 5 ECTS thesis seminar

The seminar is not embedded in DEKE0122. DEKE0121 is a separate 5 ECTS Pass/Fail course. It requires compulsory progress meetings or seminars and culminates in scientific communication around the thesis results. Do not add the seminar credits to the thesis and call the thesis 35 ECTS; the curriculum already packages the components correctly.

11. Prepare for the Young Scientists event early

DEKE0121 requires the thesis work to be communicated through an oral presentation, poster, abstract and press release, with participation in the Department of Biochemistry’s Young Scientists event or an approved alternative arrangement. The course notes that experimental work should already be underway in January for the normal spring event. This creates a real scheduling dependency.

12. Treat seminar outputs as a quality-control layer

A poster and short oral talk force you to compress the research question, method, key result and conclusion. Use that constraint as a diagnostic tool. If the main result cannot be stated without five caveats, the design or interpretation may still be unclear. Seminar feedback should feed back into the thesis, not remain a separate performance task.

13. TTDK1308 is 0 ECTS but still part of the route

For the Master’s degree, the current maturity-test course says the thesis abstract or another suitable part normally functions as the degree qualifying examination and demonstrates familiarity with the thesis field. Specific language-education circumstances can require a separate written maturity test. Resolve the applicable route from the current course and your own study history.

14. The programme’s research environment is analytically diverse

Food Development is not one method. The current curriculum and Food Sciences Unit span chromatography, mass spectrometry, NMR, targeted and untargeted metabolomics, sensory science, food chemistry, fermentation and bioprocessing, product development, food safety, in vitro models and some clinical intervention research. A good thesis uses only the methods needed for its question.

15. For chromatography, define the complete measurement chain

Document sample collection, storage, preparation, extraction, standards, calibration, blanks, quality-control samples, instrument sequence, peak processing and final calculation. Instrument response is not automatically analyte concentration. If recovery, matrix effects, carryover or drift can affect the result, show how they were assessed or controlled.

16. Method validation must match the thesis claim

An optimized UHPLC method can be a valid thesis topic, but “the method worked” is not enough. Decide which performance characteristics matter for the intended use: selectivity, linearity, precision, recovery, sensitivity, robustness or another justified set. Do not claim validated fitness for a purpose you did not test.

17. Separate technical replicates from independent experimental units

Repeated injections of one extract can estimate instrumental or analytical repeatability. They do not create new independent food samples. Likewise, several measurements from one fermentation vessel do not automatically become several independent fermentation batches. Define the experimental unit before choosing the statistical model.

18. Mass spectrometry needs identification discipline

FOOD1031 and the programme’s research environment use chromatography-MS combinations for food analysis. Preserve traceability from raw file to processed peak table. Distinguish targeted quantification, tentative annotation and confirmed identification. A plausible retention time or accurate mass alone may not justify the strongest compound-identity claim.

19. Metabolomics requires a frozen preprocessing story

DEKE0116 covers MS- and NMR-based metabolomics, data processing and metabolite identification. Record feature detection, alignment, filtering, missing-value handling, normalization, transformation and scaling. If preprocessing changes after looking at group separation, document the change and its rationale instead of presenting it as if it were pre-specified.

20. Use PCA and supervised models for the questions they can answer

PCA is useful for unsupervised structure exploration, outliers and dominant variance patterns. It does not by itself prove class discrimination. Supervised approaches such as PLS-DA can overfit, especially when features greatly outnumber samples. Separate model fitting from validation and avoid reporting training-set separation as predictive performance.

21. Handle many-feature testing explicitly

Metabolomics can generate hundreds or thousands of candidate features. Repeating ordinary significance tests without accounting for multiplicity can produce many chance findings. State the inferential strategy, correction or exploratory status. A metabolite that looks important in one model is not automatically a validated biomarker.

22. Sensory science has multiple designs, not one “taste test”

DEKE0103 covers discrimination tests, descriptive analysis and acceptance tests. These answer different questions. A trained analytical panel can characterize sensory properties; a consumer sample can address liking or acceptance. Do not infer consumer preference from an analytical panel unless preference was actually measured.

23. Control the sensory serving process

When relevant, define sample coding, blinding, serving order, temperature, portion size, palate cleansing, timing and carryover control. Repeated ratings from one assessor are correlated observations. The statistical design should respect assessor and product structure rather than treating every rating as independent.

24. Sensory participants can create ethics and privacy obligations

Food research does not become non-human research merely because the stimulus is food. If you recruit people for sensory or consumer tests, resolve participant information, consent, personal-data handling and the applicable ethics route before recruitment. University guidance recommends Master’s-level designs that avoid formal ethical-review triggers where possible, but that is not a blanket exemption.

25. Fermentation studies need true batch logic

DEKE0118 develops fermented foods from raw material through final-product testing. Record inoculum, substrate, process conditions, time, temperature, sampling schedule and batch identity as relevant. Separate biological batch variation from analytical replicate variation. One successful laboratory batch does not prove industrial robustness.

26. Product development has several evidence stages

DEKE0123 follows product development from idea generation and screening through technical development, pilot scale, test production and test market. A prototype can be technically successful while still lacking evidence on shelf life, safety, legal compliance, consumer acceptance or market fit. State exactly which stage the thesis reaches.

27. Food safety claims must be endpoint-specific

DEKE0119 covers hygiene, HACCP, toxicology, allergens and European food legislation. Avoid the phrase “the product is safe” unless the evidence genuinely covers the relevant hazards and legal requirements. A microbiological result does not answer every chemical or allergen question; a composition result does not establish full legal compliance.

28. Health-effect claims require the right evidence level

The programme studies healthy foods, and FOOD1030 covers nutrition research, dietary assessment, gut microbiota and health claims. But composition, in vitro activity, sensory acceptance and human clinical effect are different layers of evidence. If the thesis measures chemistry only, write a chemistry conclusion, not a clinical conclusion.

29. Correlation between chemistry and sensory response is not automatic causation

A compound concentration may correlate with bitterness, aroma or liking, but both variables can respond to processing, cultivar or matrix effects. Use causal language only when design and evidence support it. Otherwise describe association and discuss plausible mechanisms separately.

30. Keep food-law questions current and scoped

Food legislation changes. If the thesis compares a result with a legal limit, health-claim rule or safety requirement, record the exact current source and date. Do not rely on a lecture slide or an older thesis as the final legal authority. Separate scientific interpretation from regulatory determination.

31. Build a reproducible sample lineage

Each reported result should be traceable from source sample or product batch through preparation, storage, instrument run, processing and final analysis table. Use stable identifiers. Avoid manually renaming samples in multiple spreadsheets without a controlled mapping file.

32. Preserve instrument and software context

Record instrument method versions, relevant software versions and important processing parameters. This matters in chromatography, MS, NMR and multivariate analysis because a changed integration rule, library or preprocessing setting can change the final result even when the raw file is unchanged.

33. Treat failed experiments as data about the process

FairUTU emphasizes reliability, honesty and responsibility. Failed extraction, unstable fermentation, contaminated batch or unusable sensory session may need exclusion, but the exclusion should be documented. Deleting inconvenient observations without a transparent rule can distort the scientific record.

34. Separate public thesis content from confidential project material

University rules make accepted theses public documents and state that confidential information should not be included. This matters for company-sponsored product development, proprietary formulations and unpublished industrial data. Resolve confidentiality before the work begins so the thesis can remain scientifically complete without exposing protected material.

35. Turnitin is an originality control, not a methods audit

Degree theses undergo mandatory Turnitin checking, and the Master’s check is approved through UTUGradu. A low similarity score does not validate calibration, experimental design or statistics. Conversely, legitimate technical phrases and references can create similarity. Treat the report as a source-use and integrity control that still requires expert interpretation.

36. Submit through UTUGradu and keep the final version frozen

UTUGradu manages originality checking, examination, approval, publication and archiving for higher-degree theses. Before final submission, freeze the final PDF, confirm title and metadata, verify that confidential information has been removed, and ensure tables/figures match the final analysis dataset. Keep the exact submitted version distinguishable from working drafts.

37. Do not invent a programme-specific examiner rule from another degree

The current public DEKE0122 entry establishes the thesis code, credits, grading and research format, but it does not publish a Food Development-specific examiner count or final approving officer. University-wide rules govern evaluation, while detailed faculty/department procedures may live in current internal instructions. Check those instructions rather than borrowing an examiner chain from Health Technology, Computing or another Turku programme.

38. Use a final claim audit before submission

For every conclusion, label the evidence type: analytical measurement, metabolomics model, sensory result, fermentation experiment, product-development result, safety/legal comparison, in vitro result or human-health evidence. Then ask whether the verb matches the design. “Measured,” “associated,” “preferred,” “feasible,” “compliant with the tested criterion,” and “improved” are not interchangeable.

39. A practical 12-step completion workflow

1. Confirm the current Peppi route and personal study plan. 2. Secure a feasible thesis topic and supervisor. 3. Complete DEKE0120 literature work and research plan. 4. Define experimental unit, controls and analysis before data generation. 5. Resolve instrument access, safety, ethics, permits and confidentiality. 6. Run pilot checks where needed. 7. Execute the four-month experimental/measurement phase with traceable records. 8. Analyse with reproducible scripts or documented workflows. 9. Use DEKE0121 progress meetings and seminar outputs to stress-test interpretation. 10. Write DEKE0122 using the required scientific structure and current format. 11. Complete Turnitin, TTDK1308 and UTUGradu steps. 12. Verify the final public manuscript contains no confidential or identifiable information.

40. Final pre-submission checklist

Before submission, verify that DEKE0122 is still the current 30 ECTS thesis, DEKE0120 and DEKE0121 are separately complete, the maturity-test route is correct for you, the research plan and actual methods agree or deviations are explained, raw-to-result traceability is intact, statistics use the right experimental unit, sensory or participant obligations are resolved where applicable, confidential material is absent, the current faculty formatting instructions have been followed, Turnitin is complete, and the exact final file submitted to UTUGradu is the version you intend to have examined.

Additional execution notes for the experimental record

Treat the laboratory notebook and digital analysis record as part of the thesis evidence, not as disposable working material. For every experimental day, record who prepared the sample, which batch or lot it came from, storage history, preparation steps, instrument method, sequence position, deviations and any event that could affect interpretation. If a sample is re-run, preserve the original result and document why the repeat was needed instead of silently replacing it. If a calibration fails or a QC drifts outside an acceptable range, decide in advance what constitutes an acceptable rerun, reintegration or exclusion. This makes the Results section defensible because every reported number can be connected to the conditions that produced it.

For quantitative chemistry, decide how uncertainty enters the final claim. A concentration difference can come from true sample variation, extraction variability, instrumental precision and model assumptions. Report enough information to distinguish those layers. When samples are collected across days or production batches, randomization or blocking can prevent time and batch from becoming confounded with the experimental groups. If all control samples are measured first and all treated samples last, instrumental drift can imitate a treatment effect. A balanced run order and regular QC samples are often more valuable than adding many extra technical replicates at the end.

For metabolomics, create an analysis freeze before writing the main biological interpretation. Save the exact feature matrix, sample metadata, preprocessing parameters and statistical code or software settings that produced the final figures. Check whether outliers are analytical failures, data-entry problems or real biological/product variation before removing them. If supervised classification is used, report how tuning and validation were separated. Cross-validation can still be optimistic when feature selection occurs outside the validation loop, so the complete modelling pipeline matters. If the work is exploratory, saying so clearly is scientifically stronger than presenting unstable discrimination as a finished diagnostic model.

For sensory work, define the panel before collecting ratings. Record whether assessors are trained, semi-trained or consumers; inclusion criteria; approximate target population; number of assessors; repeated sessions; and whether the same products are assessed multiple times. Plan the serving order and coding scheme so brand, treatment or expected quality is not obvious. Consider fatigue, adaptation and carryover when many samples are served. When the design uses repeated measures, assessor effects and product-by-assessor variability may need to be represented in the analysis rather than collapsed into a single mean.

For fermentation and product-development experiments, keep formulation and process versions under simple version control. A small change in starter culture, water activity, pH target, heating profile, mixing time, storage condition or packaging can change the outcome. The thesis should make clear which version generated which result. If the final prototype is chosen after several iterations, describe the decision criteria and do not report only the successful final version as if no alternatives were tested. Negative iterations can explain why the final process was selected and improve the technical credibility of the discussion.

Finally, distinguish internal project decisions from formal university requirements. A company may impose a confidentiality agreement, a research group may have an instrument booking rule, and a supervisor may prefer a particular reference style. These can be important operational constraints, but they are not automatically university-wide rules. In the final thesis, identify which instructions come from the current DEKE course, which come from Faculty or University processes, and which are local project arrangements. That separation reduces confusion when the thesis moves from laboratory work to examination, publication and archiving.