STREGA Checklist: 22 Items Applied to a Real Open-Access Genetic Association Study

Published November 1, 2026

The STREGA checklist is the reporting guideline for genetic association studies. The name stands for STrengthening the REporting of Genetic Association studies, and Little and colleagues published it in 2009 as an extension of the STROBE statement ^[1]. STROBE has 22 items, and STREGA adds genetics-specific text to 12 of them ^[1]. The additions deal with problems that rarely arise in other observational designs, such as genotyping error, population stratification, Hardy-Weinberg equilibrium and the large number of variants that can be tested.

Most guides stop at listing the items. This guide goes further. After the checklist itself, we apply every STREGA row to a real open-access case-control study, published in the journal Hormones in 2026, that states it followed STREGA, with the article section beside every verdict ^[2]. The need is documented. Nedovic and colleagues examined 160 genetics and heredity journals and found that only 29 (18.1%) referenced STREGA in their instructions for authors, and adherence to the checklist was significantly higher in journals that endorsed it ^[3].

What the STREGA Checklist Is and How It Relates to STROBE

STREGA is the genetic counterpart of the STROBE guideline for observational studies, which our STROBE guide explains in full. A multidisciplinary group built it from a literature review, workshop presentations and discussion, and iterative electronic correspondence. Thirty-three of 74 invitees attended a workshop in Ottawa in June 2006, among them epidemiologists, geneticists, statisticians, journal editors and graduate students ^[1]. The statement appeared in 2009 in several journals, and the PLoS Medicine version is released under a Creative Commons CC0 dedication, so its wording can be reused freely ^[1].

The authors say their recommendations are not meant to support or oppose any particular study design or method. The aim is transparency and completeness in reporting what was done and found ^[1]. A checklist of this kind therefore records whether a reader can find the information, and it does not rate how good the study was. The reporting gap it targets was visible early. The authors cite an assessment of a random sample of 315 genetic association studies published from 2001 to 2003, in which only some studies described methods to validate genotyping ^[1].

STREGA Checklist Items by Manuscript Section

Table 1 paraphrases the 22 items and shows what STREGA adds to each, grouped by the section of the manuscript where the answer belongs. The wording is ours, and the original text sits in Table 1 of the open-access paper ^[1]. By our count the additions amount to 18 extra or extended lines across 12 items, with the heaviest load on item 12, the statistical methods.

Table 1. The 22 STROBE items and the STREGA additions, paraphrased from Little et al., 2009 (CC0).

Figure 1 shows the same structure at a glance. Twelve of the 22 boxes carry a genetics addition, and they cluster in Methods and Results, where the laboratory, the statistics and the participant counts live. Introduction and Discussion items, with the exception of item 3, keep their STROBE wording.

The Five Genetic Problems STREGA Targets

The STREGA authors single out five areas of special interest that are specific to, or especially relevant in, genetic association studies ^[1]. Table 2 sets out each one with the item that handles it and the reason it matters.

Table 2. The five main areas of special interest in STREGA, paraphrased from Little et al., 2009.

The authors list seven further issues, namely selection of participants, the rationale for choosing genes and variants, treatment effects in studies of quantitative traits, statistical methods, relatedness, reporting of descriptive and outcome data, and the volume of data tested ^[1]. The last of these matters most for genome-wide work, where a search across hundreds of thousands of variants can be expected to produce many false-positive odds ratios by chance alone. Our guide to statistical methods in genomics covers the testing side, and our chi-square guide shows the test that compares genotype counts and checks for equilibrium departures.

Choosing Between the STREGA Checklist, STROBE and STROBE-MR

Use the STREGA checklist when the paper reports an association between a genetic variant and a trait or disease in a cohort, case-control or cross-sectional design, whether the variants were picked as candidates or scanned genome-wide. The STREGA table includes the STROBE wording, so one completed table covers both. A Mendelian randomization paper is different, because it treats genetic variants as instruments to estimate a causal effect, and it has its own 20-item guideline, STROBE-MR ^[5]. Our STROBE-MR worked example scores a real study item by item in the same way as this guide.

For the method itself, start with our explainer on Mendelian randomization, then read how the F-statistic flags weak instruments, what each of the four estimators protects against in our pleiotropy guide, and why the power calculation so often decides a desk rejection. If you are unsure which guideline fits your design, use the design named in the title as the guide, with STREGA for genetic association studies and STROBE-MR for Mendelian randomization. Reviewers who pool genetic association studies may also rate case-control studies with the Newcastle-Ottawa Scale, which is a quality rating tool and not a reporting guideline, so the two answer different questions.

The Open-Access Example We Scored

The article we scored

We chose Kalesis and colleagues, published in Hormones in 2026 on the genetic background of obesity in a case-control design ^[2]. The paper was received on 7 October 2025, accepted on 24 April 2026 and published online on 2 June 2026. The abstract states “The study was conducted and reported in accordance with STREGA guidelines.”, and the article cites Little and colleagues as its reference for the guideline. The study compares 257 Greek adults with obesity, 136 without type 2 diabetes and 121 with it, at three variants, TCF7L2 rs7903146, FTO rs9939609 and PPARG rs1801282, and it adds an HbA1c analysis within the non-diabetic group.

The licence is CC BY 4.0, and three independent records agree. The Crossref licence record for the version of record gives the Creative Commons Attribution 4.0 URL with a start date of 2 June 2026, the Europe PMC licence field reads “cc by”, and the licence statement inside the article XML (JATS) names the Creative Commons Attribution 4.0 International License. The PMC open-data record also lists the licence as CC BY and the article as not retracted. Because a PMC copy can differ from the publisher version, we compared the XML body text with the typeset PDF held in the PMC open-data bucket, which carries the journal’s own layout and page numbers 527 to 533, and found no differences in the Methods, Results or Discussion wording. Page pointers below refer to that PDF.

We found no completed STREGA checklist among the article files, which are the text, the figure, the XML and the PDF. The only figure is a forest plot of odds ratios, so we reproduce no figure and the tables are summarised in our own words.

How we scored it

The STREGA table has 22 items, and several have lettered parts, so we scored 49 rows, each with the article section or page beside it. Rows that STREGA adds to STROBE are tagged. Each row carries one of five verdicts. Reported means the information is present and clear. Partly reported means some of it is present, or it is present but inconsistent. Not reported means we found nothing. Not applicable means the row does not fit the design, and we say why. Judgement call means reasonable readers could score the row either way, and we explain our reading.

Two cautions apply. First, STREGA is a reporting guideline, so a row marked Not reported means we could not find the information in the article, and the authors may hold it elsewhere. Second, we score only what the published article shows. Quotations are verbatim, no longer than 25 words, and checked against the typeset PDF.

STREGA Checklist Applied to the Example, Item by Item

Table 3. STREGA applied to Kalesis et al., Hormones 2026 (CC BY 4.0). Page numbers and table references in the third column refer to the article, Hormones 2026, volume 25, issue 2, pages 527 to 533.

What the Tally and the Numbers in the Example Show

Across 49 rows the tally is 17 Reported, 15 Partly reported, 5 Not reported, 6 Not applicable and 6 Judgement call. Of the 18 rows that STREGA adds, 6 are Reported, 4 Partly reported, 3 Not reported, 2 Not applicable and 3 Judgement call. The counts are for this one paper and are not a quality score, and a different reader would move some Judgement call rows.

The strengths sit in the laboratory and in the full reporting of results. The article names the genotyping system and laboratory, gives a 100% call rate and 100% duplicate concordance, reports exact Hardy-Weinberg tests for every variant, and shows every variant with every genetic model. The gaps sit in places where STREGA is most specific, namely the study size, multiple comparisons, relatedness, population stratification, and the stage counts before genotyping.

Arithmetic we checked

We recomputed the main numbers from the printed tables, using the control and case genotype counts as printed. Table 4 below sets the printed values beside our results. The agreement is close on the core results, and it breaks in one place.

Table 4. Numbers in the example recomputed from the printed tables. The arithmetic is ours.

The article Table 5 result needs care. The printed counts reproduce the printed p values, so the discrepancy sits either in the column headings or in the counts for two of the three variants. Genotyping was complete for all 257 participants, so the same people should fill the same two HbA1c groups for every variant, and the article alone does not let us tell which figure is intended. We note it as a point to reconcile and not as an error of method.

Inconsistencies a reader will notice

• Article Table 4 labels three rows in its Notes column as “Primary model; unadjusted”, namely the TCF7L2 recessive, FTO dominant and PPARG recessive rows, while its footnote says “The additive (per-allele) model is the prespecified primary analysis; dominant and recessive models are secondary”. As printed, the labels contradict the footnote and the Methods.
• Article Table 5 group totals do not agree across the three variants, as set out above.
• The Methods sentence “The revised report was prepared in accordance with STREGA guidance” carries reference number 19, which in the reference list is the American Diabetes Association Standards of Care, while the STREGA paper is reference 15.
• The Discussion says the quality control data “support retaining these loci in an exploratory analysis”, and the Limitations say Hardy-Weinberg departures at all three variants raise “concerns regarding genotyping validity, sampling effects, or population structure”. Both statements are open and honest, and a reader has to weigh them together when reading the odds ratios.

Five Reporting Lessons From the Example

Describe the laboratory the way item 8b asks

The example is strong on quality control, with a named platform, a named laboratory, a 100% call rate and duplicate genotyping of about 10% of samples. It still leaves out the calling software and version, the storage of DNA and any statement on batches or on whether cases and controls were genotyped together. These are short sentences that STREGA asks for because differential misclassification between cases and controls can arise from differences in DNA storage, collection or processing even when the genotyping itself is excellent ^[1].

Decide what Hardy-Weinberg departures mean before the Discussion

Item 12f asks whether equilibrium was considered and how, and the example answers that clearly with an exact test ^[4]. The STREGA authors add that any procedure to allow for deviations should also be described, and that for large platforms the threshold used to exclude variants should be stated ^[1]. In the example all three variants depart from equilibrium in controls, and the heterozygote deficits are marked for two of them. A reader needs to see in Methods what would have been done in that event, for example re-checking the assay or running the analysis with and without the variant, because the Discussion then has to carry the whole interpretation.

Say something about multiplicity and stratification, even if the answer is none

Rows 12h, 12i and 16d are the ones where a one-line statement costs the authors almost nothing. Three variants and three genetic models give nine association estimates, and the article states neither a correction nor a reason for none. Stratification is handled by restricting to one ancestry according to the study record, which is a legitimate choice that the STREGA authors would want stated together with how ancestry was ascertained. A sentence saying that no correction and no stratification method was applied, and why, satisfies the item.

Give the stage counts and the basis for the study size

The article gives final group sizes and the genotyping success rate, and not the numbers approached, eligible or excluded, or how 257 was chosen. Those three facts let a reader judge selection and power, and they are the usual contents of a flow diagram. Our free flow diagram generator, listed below, draws a STROBE diagram from the counts.

Make every label, total and citation number agree

The few real inconsistencies in the example are small ones, namely Notes labels in one table, group totals in another and a citation number that points to the wrong reference. None changes the odds ratios, and each is the kind of detail that editors and reviewers check first. A consistency pass, with every count, label and reference number compared between the abstract, the tables, the text and the reference list, takes under an hour.

How to Complete Your Own STREGA Checklist

Start from the STREGA table in the open-access paper, which the authors also provide as an editable file, Table S1 in the supporting information ^[1]. Add one row for each item, write the manuscript section or page beside it, and write a reason in every row that does not apply. Then write the genetics paragraphs below, each short enough to sit in Methods or Results.

• Laboratory paragraph. Source and storage of DNA, genotyping method and platform, calling algorithm and version, laboratory, call and error rates, comparability across groups, and whether genotypes were assigned together or in batches.
• Hardy-Weinberg sentence. The test used, the group it was applied to, any threshold for excluding variants, the result for each variant, and what you did when a variant departed.
• Stratification sentence. The ancestry information available, how it was ascertained, and the method used to assess or address stratification, or a statement that none was used.
• Multiplicity sentence. How many variants and models were tested, and whether and how the threshold was adjusted.
• Counts. Numbers genotyping was attempted and succeeded, numbers per genotype category for cases and controls, and the stages before them.
• Replication statement. Whether the study is a first report, a replication or both, with the criteria for replication when relevant.

Authors who need the design guideline first can use our STROBE guide, and the official STREGA page is listed on the EQUATOR Network site. The original STROBE statement is also open access ^[6], and the STROBE website carries the checklists for each design (STROBE statement site).

Common mistakes with the STREGA checklist

• Writing that the study “followed STREGA” without supplying the checklist or the item locations.
• Reporting a call rate without the method, the platform or the laboratory.
• Testing Hardy-Weinberg equilibrium and then saying nothing about the result or what it meant for the analysis.
• Testing many variants or models without a statement on multiplicity.
• Giving only final group sizes and leaving out the numbers approached, excluded and genotyped.
• Leaving table labels, group totals and citation numbers out of step with the text.

Professional Support for a Genetic Association Manuscript

A checklist records what the study already contains, so the planning stage matters as much as the writing stage. Questions such as how cases and controls are selected, how the sample is drawn and how the laboratory work will be reported are easier to settle before data collection. Two Editverse services map onto the two stages. One is a methodology consultation for the design. The other is substantive editing of the finished manuscript with verification of the reference list. Editverse editing and writing support aligns with ICMJE, COPE and GPP 2022, and with the ethics requirements of your institution and country ^[7][8].

If your question is a causal one that public GWAS summary statistics can answer, the Mendelian randomization study package includes a STROBE-MR checklist.

Conclusion

The STREGA checklist keeps all 22 STROBE items and adds genetics text to 12 of them, so that a reader can judge how DNA was handled, how many samples were genotyped, whether genotype counts fit equilibrium, how ancestry was treated and how many variants were tested. Applied to a real open-access case-control study, it shows what good reporting looks like, with a named platform and laboratory, a full call rate, duplicate genotyping and every variant shown under every model. It also shows where even a careful paper leaves gaps, namely study size, multiplicity, relatedness, stratification and stage counts, and where small inconsistencies creep in. Complete the table before submission, write a reason in every row that does not apply, and check every count, label and citation number in the final proofs.

Frequently Asked Questions About the STREGA Checklist

How many items does the STREGA checklist have?

It keeps the 22 items of the STROBE checklist and adds genetics-specific text to 12 of them, according to the 2009 statement. By our count the additions run to 18 lines, including the new sub-items 12f to 12j, 16d, 17b and 17c.

Does the STREGA checklist replace STROBE?

No. It is an extension, and the STREGA table contains the STROBE wording alongside the genetics additions, so one completed table covers both.

Does the STREGA checklist apply to genome-wide association studies?

Yes. The authors address genome-wide studies directly, for example by asking for the thresholds used to exclude variants and for a summary of all variants examined when many were tested.

Is the STREGA checklist used for Mendelian randomization?

No. Mendelian randomization studies have their own guideline, STROBE-MR, with 20 items.

Is the STREGA checklist a quality score?

No. It records whether the reporting is complete, and it does not rate the quality of the study. Counts of reported and unreported rows describe one paper and should not be turned into a scale.

Where can the official STREGA checklist be found?

The statement is open access in PLoS Medicine under a CC0 dedication, with an editable version of the table in the supporting information, and it is listed on the EQUATOR Network site.

What did the example paper leave out most often?

In our worked example the clearest gaps were the basis for the study size, multiple comparisons, relatedness, the stage counts before genotyping and the handling of population stratification.

References

1. Little J, Higgins JPT, Ioannidis JPA, Moher D, Gagnon F, von Elm E, et al. STrengthening the REporting of Genetic Association Studies (STREGA): an extension of the STROBE statement. PLoS Med. 2009;6(2):e1000022. doi:10.1371/journal.pmed.1000022. PMID: 19192942. PMCID: PMC2634792. CC0. Publisher page.
2. Kalesis K, Roidos P, Nikolopoulos V, Katsiki N, Kotsa K, Koufakis T. Genetic background of obesity: a case-control study. Hormones (Athens). 2026;25(2):527-33. doi:10.1007/s42000-026-00790-3. PMID: 42230483. PMCID: PMC13582106. CC BY 4.0. Publisher page.
3. Nedovic D, Panic N, Pastorino R, Ricciardi W, Boccia S. Evaluation of the endorsement of the STrengthening the REporting of Genetic Association Studies (STREGA) statement on the reporting quality of published genetic association studies. J Epidemiol. 2016;26(8):399-404. doi:10.2188/jea.JE20150173. PMID: 27349199. PMCID: PMC4967660. Facts only are cited. Publisher page.
4. Wigginton JE, Cutler DJ, Abecasis GR. A note on exact tests of Hardy-Weinberg equilibrium. Am J Hum Genet. 2005;76(5):887-93. doi:10.1086/429864. Publisher page.
5. Skrivankova VW, Richmond RC, Woolf BAR, Yarmolinsky J, Davies NM, Swanson SA, et al. Strengthening the reporting of observational studies in epidemiology using Mendelian randomization (STROBE-MR). JAMA. 2021;326(16):1614-21. doi:10.1001/jama.2021.18236.
6. von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Med. 2007;4(10):e296. doi:10.1371/journal.pmed.0040296.
7. International Committee of Medical Journal Editors. Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals. Available at icmje.org/recommendations.
8. DeTora LM, Toroser D, Sykes A, Vanderlinden C, Plunkett FJ, Lane T, et al. Good Publication Practice (GPP) guidelines for company-sponsored biomedical research, 2022 update. Ann Intern Med. 2022;175(9):1298-1304. doi:10.7326/M22-1460. PMID: 36037471.