A summary of published research, with citations. It states no amount, schedule or route, and makes no claim about what this compound does in a person. Supplied for laboratory research use only.
On this page · 7 sections
No gene encoding DSIP (delta sleep-inducing peptide) has been identified, no receptor for it has been cloned, and no study has tested whether any genotype changes the response to it. A consumer sleep-genetics report can describe variants linked to chronotype, deep-sleep intensity or insomnia risk, but none of those variants has ever been studied alongside DSIP. The accurate answer to "what do my genes say about DSIP?" is that no study has ever asked the question.

That answer is less satisfying than a report with coloured bars, but it is the one the literature supports. The rest of this page sets out why, starting with the gene that is missing.
Is there a gene that encodes DSIP?
Not one that anyone has found. DSIP is a nine-residue peptide, Trp-Ala-Gly-Gly-Asp-Ala-Ser-Gly-Glu, characterised in 1977 from the cerebral venous blood of rabbits held in an electrically induced sleep state (Schoenenberger and Monnier, PMID 265572). The history of that discovery and the sleep studies that followed is covered separately; what matters here is what happened next, which is that nobody found where the peptide comes from.
Most neuropeptides are cut out of a larger precursor protein, and the precursor is encoded by a gene that can be sequenced, knocked out and genotyped. ACTH and the melanocortins come from pro-opiomelanocortin; oxytocin and vasopressin each have their own precursor gene. For DSIP, no such precursor has been identified. Kovalzon and Strekalova, reviewing the field in 2006, put the lack of an isolated DSIP gene, protein and receptor at the centre of why the sleep hypothesis remained "extremely poorly documented and still weak" (PMID 16539679).
The closest anyone has come is a database search. Mikhaleva and colleagues scanned mammalian gene and protein databases for stretches resembling DSIP and found, inside a family of histone demethylases encoded by the JMJD1B gene, the sequence WKGGNASGE, which differs from DSIP at two of its nine positions (PMID 21262293). They proposed these proteins as possible precursors of peptides with DSIP-like activity. Note the wording: a related sequence, a proposal, and a preliminary test of the synthetic fragment in rodent behavioural assays. The exact DSIP sequence was not found in a precursor, and the authors state plainly that nothing was known of its precursor proteins or their genes.
The gene that carries DSIP in its name
A search of a raw genotype file for "DSIP" can turn up a gene called DSIP immunoreactor. The name causes confusion, so it is worth explaining.
In 1993 Sillard and colleagues purified a 77-residue peptide from porcine brain that was recognised by an antiserum raised against synthetic DSIP. Its amino acid sequence, they reported, "is not related to that of DSIP"; it carried a leucine-zipper motif instead (PMID 8375381). The antibody had cross-reacted with a different molecule. The gene later catalogued as the DSIP immunoreactor is better known as glucocorticoid-induced leucine zipper, or GILZ, a regulator of transcription studied in immune and adipose tissue (Gimble and colleagues, PMID 19849801).
So the one gene with DSIP in its name records an antibody cross-reaction, not ancestry. A variant in it tells you nothing about DSIP.
What consumer sleep-genetics tests actually report
The variants that appear in sleep reports are real, and some are well studied. None was identified in, or has been tested in, a DSIP experiment.
| What the report shows | What the published work found | Connection to DSIP |
|---|---|---|
| Chronotype (morning or evening type) | 351 loci across 697,828 people; extreme polygenic groups differ by about 25 minutes in average sleep timing | None studied |
| PER3 repeat length | Associations with diurnal preference and responses to sleep loss and light; several studies small, some unreplicated | None studied |
| ADA variant | Variant with reduced adenosine breakdown associated with more deep sleep and slow-wave activity | None studied |
| ADORA2A variant | Linked to caffeine-related anxiety; altered EEG in sleep and wakefulness alike | None studied |
| DEC2 (BHLHE41) mutation | One mutation found in people with a habitual short-sleep phenotype, modelled in mice | None studied |
| Insomnia risk score | 554 risk loci in a meta-analysis of more than 2.3 million people | None studied |
The chronotype result is the most useful one for calibrating expectations. Jones and colleagues found that the 5% of people carrying the most morningness alleles slept, on average, 25 minutes earlier than the 5% carrying the fewest (PMID 30696823). That is the gap between the two extreme tails of hundreds of thousands of genotyped people. A single person's score sits somewhere in the middle and moves the estimate by less.
The insomnia picture is similar. Watanabe and colleagues' meta-analysis of 593,724 cases and 1,771,286 controls identified 554 risk loci, and the authors described the trait's architecture as extremely polygenic (PMID 35835914). A trait spread thinly across hundreds of loci cannot be read off any single variant.
The single-gene findings are more striking and narrower. Rétey and colleagues reported that a functional variant of adenosine deaminase, associated with slower conversion of adenosine to inosine, went with longer and more intense deep sleep in their participants, while an ADORA2A polymorphism affected the EEG in sleep and wakefulness without being specific to either (PMID 16221767). He and colleagues identified a DEC2 mutation associated with a human short-sleep phenotype and showed that transgenic mice carrying it slept less (PMID 19679812). Archer and colleagues' review of PER3 found associations across diurnal preference, light responses and cognitive responses to sleep loss, while noting that several studies were small and some findings were not replicated (PMID 29248294).
Every one of those papers is about how sleep varies between people. None is about how people respond to a compound, and none involves DSIP.
Why no genotype can predict a response to DSIP
Predicting response from DNA is called pharmacogenetics, and it works only when several links are in place. The cases that hold up in practice share the same chain: a known target or metabolising enzyme, a gene variant that changes that protein, a measurable difference in exposure or target activity, and controlled studies showing that people with different genotypes respond differently on an outcome that matters.
DSIP has none of those links.
There is no identified target. Binding sites were reported on cultured neurones in the 1980s, but no receptor has been cloned, so there is no receptor gene in which a variant could sit.
There is no characterised pathway to genotype. Without a known precursor or receptor, there is no candidate gene to examine at all, let alone one whose variants are known to change the handling of the peptide.
There are no genotype-by-treatment studies. The human DSIP studies date from 1981 to 1992, predate routine genotyping, and enrolled participants in the tens (Bes and colleagues, PMID 1299794, is a representative double-blind example). No study has stratified participants by any variant.
Association is not modification. Even if a variant is reliably linked to a sleep trait, that does not make it a modifier of a compound's effect. An ADA carrier's deeper baseline sleep says nothing about whether a peptide with an unknown mechanism would act differently in them. Showing that a genotype modifies a response is a test of an interaction, and interactions need far larger samples to detect than the main effects that sleep-genetics studies report.
A report that combines your PER3 result with a DSIP recommendation has joined two things that the literature has never connected.
What the DSIP literature actually examined
The DSIP literature is mostly animal electroencephalography, endocrine and stress measurements from the 1970s and 1980s, plus a handful of small human sleep studies. Its own reviewers did not find it settled. Kovalzon and Strekalova reported that in their earlier work, certain synthetic structural analogues of DSIP promoted slow-wave sleep in rabbits and rats while DSIP itself did not, and they proposed that a still-unidentified DSIP-like peptide might account for both the immunoreactivity and the reported activity (PMID 16539679).
That is the state of the question: an unresolved molecule, no confirmed gene, and a hypothesis that the real actor may be something else. For how to weigh small, old and unreplicated studies like these, see how to read a peptide study. For why a missing receptor matters so much, see how peptide receptors work. And for a neighbouring literature on circadian signalling, Epithalon, melatonin and the pineal gland has the same single-group limits.
Where the evidence stands
Evidence stage. Animal studies and a small, old set of human sleep studies. No modern randomised trial. A search of ClinicalTrials.gov for delta sleep-inducing peptide returns no registered trial of the peptide.
Regulatory status. DSIP is not an approved medicine in Canada and has not been approved by the FDA or the EMA for any indication. Health Canada's position on research peptides generally is covered in what Health Canada's rules mean for research peptides.
Genetics. No DSIP gene, no DSIP receptor gene, and no genotype-by-DSIP study. Sleep-genetics variants describe traits, not responses to this compound.
Nothing here is medical, dosing, or therapeutic guidance.
Supplied here for laboratory use: DSIP. The laboratory's certificate for each lot is linked from its product page. For laboratory research use only.
Frequently asked questions
Is there a DSIP gene I can look up in my raw DNA data? No gene encoding DSIP has been identified. A search may return DSIP immunoreactor, also called GILZ, but that gene was named because its protein product was recognised by an antibody against DSIP. Its sequence is unrelated to DSIP, so a variant there tells you nothing about the peptide.
Can a consumer DNA sleep report tell me whether DSIP would work for me? No. Those reports describe variants associated with chronotype, deep-sleep intensity or insomnia risk across large populations. None of those variants has been tested in any DSIP study, and no study has compared DSIP responses between genotypes. A report linking the two is making a connection no published research supports.
Does carrying a short-sleep or night-owl variant change how DSIP behaves? Unknown, and untested. Variants such as the DEC2 short-sleep mutation or PER3 repeat lengths are linked to how sleep varies between people. Whether any variant changes a response to DSIP would need a genotype-stratified controlled trial, and none has been done.
Why has nobody found the DSIP gene after fifty years? The peptide's origin has never been traced to a precursor protein, and no receptor has been cloned. The closest lead is a related nine-residue sequence inside the JMJD1B histone demethylase, which differs from DSIP at two positions. Reviewers have also proposed that a different, unidentified DSIP-like peptide may explain the earlier findings.
Is DSIP an approved sleep treatment anywhere? No. It has not been approved by Health Canada, the FDA or the EMA for any indication. Its human sleep studies are small, date from 1981 to 1992 and have not been replicated with modern methods. Material sold here is a research chemical for laboratory use only.
Selected Research References
- Schoenenberger GA, Monnier M. Characterization of a delta-electroencephalogram (-sleep)-inducing peptide. Proceedings of the National Academy of Sciences of the United States of America. 1977;74(3):1282-1286. pubmed.ncbi.nlm.nih.gov/265572
- Kovalzon VM, Strekalova TV. Delta sleep-inducing peptide (DSIP): a still unresolved riddle. Journal of Neurochemistry. 2006;97(2):303-309. pubmed.ncbi.nlm.nih.gov/16539679
- Mikhaleva II, Prudchenko IA, Ivanov VT, et al. JmjC-domain-containing histone demethylases of the JMJD1B type as putative precursors of endogenous DSIP. Peptides. 2011;32(4):826-831. pubmed.ncbi.nlm.nih.gov/21262293
- Sillard R, Schulz-Knappe P, Vogel P, et al. A novel 77-residue peptide from porcine brain contains a leucine-zipper motif and is recognized by an antiserum to delta-sleep-inducing peptide. European Journal of Biochemistry. 1993;216(2):429-436. pubmed.ncbi.nlm.nih.gov/8375381
- Gimble JM, Ptitsyn AA, Goh BC, et al. Delta sleep-inducing peptide and glucocorticoid-induced leucine zipper: potential links between circadian mechanisms and obesity? Obesity Reviews. 2009;10 Suppl 2:46-51. pubmed.ncbi.nlm.nih.gov/19849801
- Jones SE, Lane JM, Wood AR, et al. Genome-wide association analyses of chronotype in 697,828 individuals provides insights into circadian rhythms. Nature Communications. 2019;10(1):343. pubmed.ncbi.nlm.nih.gov/30696823
- Archer SN, Schmidt C, Vandewalle G, Dijk DJ. Phenotyping of PER3 variants reveals widespread effects on circadian preference, sleep regulation, and health. Sleep Medicine Reviews. 2018;40:109-126. pubmed.ncbi.nlm.nih.gov/29248294
- Rétey JV, Adam M, Honegger E, et al. A functional genetic variation of adenosine deaminase affects the duration and intensity of deep sleep in humans. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(43):15676-15681. pubmed.ncbi.nlm.nih.gov/16221767
- He Y, Jones CR, Fujiki N, et al. The transcriptional repressor DEC2 regulates sleep length in mammals. Science. 2009;325(5942):866-870. pubmed.ncbi.nlm.nih.gov/19679812
- Watanabe K, Jansen PR, Savage JE, et al. Genome-wide meta-analysis of insomnia prioritizes genes associated with metabolic and psychiatric pathways. Nature Genetics. 2022;54(8):1125-1132. pubmed.ncbi.nlm.nih.gov/35835914
All products sold by NovaEvum are for research and laboratory use only. They are not intended for human or animal consumption, diagnosis, treatment, cure, or prevention of any disease.
