concurring:
I concur in the majority opinion, but write separately for two reasons. First, I seek to stress the narrowness of our holding, identifying several issues we do not decide today. Second, I wish to state my basis for agreeing that Friedman v. Boucher, 580 F.3d 847 (9th Cir.2009), does not control the outcome of this appeal.
I
A
1
In analyzing the interests of Pool at stake in this litigation, the majority opinion *1229properly distinguishes between the physical intrusion of a buccal swab and the more amorphous — and more serious — invasion of privacy caused by governmental collection of information contained in DNA samples and profiles. In addition, we must distinguish between the collection of a DNA sample and use of a DNA profile.
An individual’s genome — the full complement of DNA contained in each cell of his body — consists of several billion base pairs of DNA. Henry T. Greely et al., Family Ties: The Use of DNA Offender Databases to Catch Offenders’ Kin, 34 J.L. Med. & Ethics 248, 249 (2006). As the majority opinion notes, DNA profiles used in the government’s Combined DNA Index System (“CODIS”) do not contain information about an individual’s entire genome; the profiles contain only information about short tandem repeat (“STR”) occurring at thirteen different points, “loci,” in the genome. (Majority Op. 1215 (citing United States v. Kincade, 379 F.3d 813, 818-19 (9th Cir.2004) (en banc)).)
These STRs are stretches of DNA for which the DNA-replieation process appears to “stutter,” resulting in repeated iterations of a specific sequence of base pairs. See Greely, supra, at 249. For example, at a specific locus on a chromosome, a particular series of base pairs may appear three times in a row for some people, but five times in a row for others. Each variation in the number of repeats is identified as a different allele. See id. at 250. For each of the thirteen loci used in the CODIS database, between seven and twenty-three alleles appear in significant numbers among the population. Id. The DNA profiles contained in the CODIS database consist of a series of numbers corresponding to the length of STRs at each of the selected loci. Id. The number of STRs evaluated for an individual’s DNA profile represents a small fraction of the STRs throughout a person’s genome. Id. at 249-50.
Although the biological purpose, if any, of STRs remains debated, id., STRs are not “genes”: Unlike genes, STRs have not been shown to code for specific molecules of ribonucleic acid (“RNA”).1 Id.; see also Kincade, 379 F.3d at 818 (“These STR loci are each found on ... non-genic stretches of DNA not presently recognized as being responsible for trait coding ....); Kaye, supra, at 54.
These DNA profiles differ significantly from DNA samples. The term “DNA sample” is defined by statute as “tissue, fluid, or other bodily sample of an individual on which a DNA analysis can be carried out.” 42 U.S.C. § 14135a(e)(l). That is, the sample is made up of cells from an individual which contain that person’s entire genome. If fully analyzed, a sample would yield far more information than that contained in a CODIS profile, including information about all of the trait-coding DNA in the individual’s genome — that is, the precise content of each of her genes.
2
Despite Pool’s protestations to the contrary, at present CODIS DNA profiles are essentially useless for all but identification purposes. In this respect, they are quite similar to the information gained from fin*1230gerprinting and photographing — routine booking procedures.
Pool cites to several articles suggesting that so-called “junk DNA”2 may actually play an important biological function, including affecting how and when genes are expressed. E.g., Cole, supra note 2, at 56; Justin Gillis, Genetic Code of Mouse Published; Comparison With Human Genome Indicates ‘Junk DNA’ May be Vital, Wash. Post, Dec. 5, 2002, at Al. Yet even these articles do not support Pool’s assertion: “Although biologists are discovering functions for some types of ‘junk DNA,’ none have yet claimed that the forensic STRs do function.” Cole, supra note 2, at 56 (footnote omitted); see also Gillis, supra (surmising that in the future a small percentage of non-coding DNA “will prove to be regulatory regions”).
Even if the STRs used in CODIS contained some limited biological information, the same is true of fingerprints and photographs. As the majority opinion discusses, booking photographs might yield clues as to familial relationships. (Majority Op. 1220-21.) Such photographs necessarily contain information regarding an individual’s race, gender, and ethnic characteristics. Fingerprints, too, may correlate with certain traits. See Bert-Jaap Koops & Maurice Schellekens, Forensic DNA Phenotyping: Regulatory Issues, 9 Colum. Sci. & Tech. L.Rev. 158, 160 & n.4 (2008) (noting that fingerprint patterns may correlate with gender, homosexuality, ethnicity, and health conditions such as congenital heart disease).3
Pool’s efforts to categorically distinguish the information contained in CODIS DNA profiles from that contained in fingerprints are ultimately unpersuasive. Although the historical basis for allowing fingerprinting is not entirely clear, the near universal acceptance of the practice casts a long shadow over this case. See Napolitano v. United States, 340 F.2d 313, 314 (1st Cir. 1965) (“[Fingerprinting defendants on pretrial release] is universally standard procedure, and no violation of constitutional rights.”); United States v. Iacullo, 226 F.2d 788, 793 (7th Cir.1955) (“[Defendant’s] constitutional rights were not violated when his fingerprints were taken and at the trial used as a basis for comparison with fingerprints found on newspapers used to wrap narcotics.”). As Judge Augustus Hand opined in presaging the Court’s Fourth Amendment balancing jurisprudence by several decades: “Any re*1231straint of the person may be burdensome. But some burdens must be borne for the good of the community. The slight interference with the person involved in finger printing seems to us one which must be borne in the common interest.” United States v. Kelly, 55 F.2d 67, 68 (2d Cir.1932) (citations omitted). Pool has not provided a basis for weighing the interests in DNA profiling in a manner that is different from the interests involved in fingerprinting and photography.
Yet I stress that we do not purport to decide the hypothetical case in which a future litigant may demonstrate that CO-DIS loci do code for RNA, or that the number of repeats at CODIS loci yield information of a type unavailable in a fingerprint or a photograph; nor do we consider a case in which the nature of the genetic information stored in the CODIS database is changed from present practice.4 In such a case, a defendant’s interests could be vastly different. If that day arrives, a future court will conduct a totality-of-the-eireumstances test anew. But for now, Pool’s CODIS profile reveals only his identity, and the majority rightly factors only Pool’s interest in that identity into its Fourth Amendment balancing.
B
Although a CODIS profile contains less information than a DNA sample by several orders of magnitude, law enforcement must, of course, collect a DNA sample to create a DNA profile. These two actions — collection of a DNA sample and the creation of a DNA profile to be loaded onto the CODIS database — are squarely challenged by Pool. Our majority opinion affirms the district court’s holding that the government may collect a DNA sample from Pool that may be “used solely for criminal law enforcement identification purposes,” (Dist. Ct. Order 3), specifically, to create a DNA profile to be loaded onto the CODIS database. Having stated what the majority opinion holds, I take pains to clarify what we do not hold.
The first point, one the majority states in no uncertain terms but which bears repeating, is that this case condones DNA testing for individuals for whom a judicial or grand jury probable cause determination has been made; it does not address such sampling from mere arrestees. (See Majority Op. 1219-20, 1228.) That distinction is highly significant.5 A judicial probable cause determination limits the *1232opportunities for mischief inherent in a suspicionless search regime. As discussed further in Part II, infra, the Supreme Court has permitted some suspicionless searches when they are subject to “standardized criteria, or established routine.” Florida v. Wells, 495 U.S. 1, 4, 110 S.Ct. 1632, 109 L.Ed.2d 1 (1990) (citations omitted). However, the Court has been careful to caution that such “programmatic” searches may not be used as “a ruse for a general rummaging in order to discover incriminating evidence.” Id.
By permitting programmatic searches in the absence of particularized suspicion, we introduce a substantial danger that law enforcement personnel will use the DNA-testing regime as a pretext for obtaining evidence against individual suspects rather than as a broad-based tool for ensuring the identity of convicts and pre-trial releasees. Because of this potential for abuse, the Court has limited its approbation of programmatic searches to those “administered in good faith.” Colorado v. Bertine, 479 U.S. 367, 374, 107 S.Ct. 738, 93 L.Ed.2d 739 (1987); see also Wells, 495 U.S. at 4, 110 S.Ct. 1632. Interposing the judiciary between the executive and the citizenry provides a pre-hoc bulwark against abuse in addition to the post-hoc good faith inquiry.
The second important limitation on our holding today is that we do not consider a claim of misuse of the CODIS system. Pool focuses his argument on the potential that the government may harvest intimate and revealing information from his DNA sample. We must not underestimate these legitimate concerns. Although the CODIS database uses only DNA profiles, the government must obtain a DNA sample to create the profiles. Government possession of the entire sample, with its potential to provide much more personal information, presents an unquestionable opportunity for abuse.
Nonetheless, Pool has advanced no evidence suggesting that the government will engage in the misconduct of which he warns, and we must be cognizant of the penalties for such misconduct, see 42 U.S.C. § 14135e(c). Whenever the government conducts a search, there is a risk that it will abuse its powers. For example, a law enforcement search for a firearm in a residence pursuant to a valid warrant typically involves control of the entire residence. While in control of the home, officers could commit any conceivable number of unlawful acts. Yet the potential for abuse does not trump the issue of legality of a search. Admittedly, the potential for intrusion into the widest spectrum of human privacy is present in a DNA sample, but the nature of our legal analysis remains constant.
It might also be objected that this analogy is inapt because the government exercises control over a home for only a short period of time, but seeks to maintain Pool’s entire DNA sample permanently. This brings me to my third and final point regarding the limits of today’s majority: We do not decide here whether the government may indefinitely retain Pool’s DNA sample. Rather, we permit the government to collect Pool’s DNA sample to create a DNA profile for the CODIS database. Because Pool’s DNA has yet to be collected, we need not examine whether the government may retain his sample beyond the creation of the defendant’s profile.
In the event Pool is acquitted, or that the charges against him are otherwise dismissed, Pool may have his sample and profile expunged. See 42 U.S.C. § 14132(d)(1)(A). If he is convicted, this statutory provision would not apply; nevertheless, the majority opinion does not hold that indefinite retention is permissible. As noted in the Kincade concurrence, *1233the proposition that the government may retain an individual’s entire DNA sample forever presents unique questions. See 379 F.3d at 841-42 (Gould, J., concurring); cf. Cole, supra note 2, at 57 n. 15 (“Even those who are sanguine about the privacy threat posed by profiles are concerned about the storage of samples.”). It is of concern that § 14132(d)(1)(A) places the burden upon a former defendant with respect to removal. However, as was the case in Kincade, the legality of such a practice is not before us. Not until we are faced with a litigant whose case has been finalized and has left the penal system will these issues be decided. See Kincade, 379 F.3d at 841-42.
II
I write separately for a second reason: to expand upon the majority opinion’s discussion of Friedman v. Boucher, 580 F.3d 847 (9th Cir.2009). In addition to the differences between this case and Friedman stated in the majority opinion, I credit the fact that the DNA collection at issue here is conducted generally of all federal pretrial releasees, see 18 U.S.C. § 3142(b), (c)(1). The search in Friedman was directed at a single prisoner. See 580 F.3d at 854(concluding the search was not conducted pursuant to a statutory regime). It does appear counterintuitive that a search may be permissible because it is less particularized, but the Supreme Court’s holdings regarding “programmatic” searches compel this conclusion. See, e.g., Samson v. California, 547 U.S. 843, 855 n. 4, 126 S.Ct. 2193, 165 L.Ed.2d 250 (2006).
“Prior to Samson, the Court had never held that the totality of the circumstances was the appropriate test to apply in a suspicionless search of a conditional releasee. Thus, several courts had concluded that a suspicionless search could not be justified absent a special need (or some other exception).” United States v. Weikert, 504 F.3d 1, 9 (1st Cir.2007) (emphasis omitted); see also Kincade, 379 F.3d at 843 (Reinhardt, J., dissenting) (arguing, prior to Samson, that the Court had never approved “a programmatic search designed to produce and maintain evidence relating to ordinary criminal wrongdoing, yet conducted without any level of individualized suspicion”). This framework was upended by the Court in Samson, which held that suspicionless searches have been sanctioned both in cases involving “special needs” and in those considering “programmatic” searches. 547 U.S. at 855 n. 4, 126 S.Ct. 2193. The Samson Court further explained that it has “never held that these are the only limited circumstances in which searches absent individualized suspicion could be ‘reasonable’ under the Fourth Amendment.” Id.
The Court has provided little guidance as to the nature of a “programmatic” search. In Brigham City v. Stuart, 547 U.S. 398, 126 S.Ct. 1943, 164 L.Ed.2d 650 (2006), it described “checkpoints to combat drunk driving or drug trafficking” as fitting within the programmatic category. Id. at 405, 126 S.Ct. 1943. It also cited Wells, which held that the permissibility of an inventory search following impoundment turns on the degree to which such searches are standardized. Brigham City, 547 U.S. at 405, 126 S.Ct. 1943; see also Wells, 495 U.S. at 4, 110 S.Ct. 1632 (“Our view that standardized criteria, or established routine, must regulate the opening of containers found during inventory searches is based on the principle that an inventory search must not be a ruse for a general rummaging in order to discover incriminating evidence.” (citations omitted)); see also id. at 8, 110 S.Ct. 1632 (Brennan, J., concurring) (“Our cases clearly hold that an inventory search is reasonable under the Fourth Amendment only if it is done in accordance with standard procedures that limit the discretion of the police.” (citation omitted)); cf. Bertine, *1234479 U.S. at 374 n. 6, 107 S.Ct. 738 (“Our decisions have always adhered to the requirement that inventories be conducted according to standardized criteria.” (citation omitted)).
The dissent in Samson provides additional guidance. Criticizing the majority’s holding as an abandonment of the rule that a suspicionless search must be supported by a “special need,” Justice Stevens argued that “if individualized suspicion is to be jettisoned, it must be replaced with measures to protect against the state actor’s unfettered discretion.” Samson, 547 U.S. at 860, 126 S.Ct. 2193 (Stevens, J., dissenting). These “programmatic safeguards [must be] designed to ensure even-handedness in application.” Id.
Unlike the search held unconstitutional in Friedman, the statute challenged by Pool imposes a uniform burden on all federal pre-trial releasees. It is “programmatic” in the sense the Supreme Court appears to recognize as relevant to whether a suspicionless search is reasonable.
Obviously, the programmatic nature of a search does not mitigate its impact on individuals’ privacy rights. But programmatic searches do fulfill another “essential purpose of a warrant requirement,” namely “to protect privacy interests by assuring citizens subject to a search or seizure that such intrusions are not the random or arbitrary acts of government agents.” Skinner v. Ry. Labor Executives’ Ass’n, 489 U.S. 602, 621-22, 109 S.Ct. 1402, 103 L.Ed.2d 639 (1989). And standard practices do limit the ability of government to use DNA profiling as a means of targeting “political opponents and disfavored minorities,” Kincade, 379 F.3d at 848 (Reinhardt, J., dissenting), or harassing disfavored individuals. When combined with the legitimate government interest in ensuring Pool appears for trial and the limited intrusion upon Pool’s privacy, I am convinced that the programmatic nature of the DNA profiling program is reasonable.
Ill
This is a vexing case. The DNA profiling system at issue promises enormous potential as an investigatory tool, but its expansion or misuse poses a very real threat to our privacy. The distinctions I attempt to draw in this concurrence are dwarfed by the magnitude of these competing interests, but we must draw lines as best we can. I therefore concur, leaving for another day difficult questions regarding the administration of CODIS and the government’s retention of DNA samples.
. "Code for” is a term of art related to the process of transcription and translation. If a stretch of DNA “codes for” RNA, that stretch will produce a certain type of RNA during the process of transcription. RNA, in turn, usually codes for protein, which means that the RNA will produce a certain protein during the process of translation. See Greely, supra, at 249. These proteins can affect an individual's physical characteristics, such as eye color or propensity to develop certain diseases. See David H. Kaye, GINA’s Genotypes, 108 Mich. L.Rev. First Impressions 51, 54 (2010).
. "Junk DNA” is a term that generally refers to any non-genic DNA, that is, stretches of DNA that do not code for RNA. Simon A. Cole, Is the "Junk ” DNA Designation Bunk?, 102 Nw. U.L.Rev. Colloquy 54, 56-57 (2007). STRs are one of several types of non-genic DNA sequence. See Greely, supra, at 249.
. Pool points to one study showing some association between TH01(one of the CODIS loci) and the insulin minisatellite (an STR not included in the CODIS database), which correlates to susceptibility to polycystic ovary syndrome, obesity, and type-1 and type-2 diabetes. John D.H. Stead, Influence of Allele Lineage on the Role of the Insulin Minisatellite in Susceptibility to Type 1 Diabetes, 9 Human Molecular Genetics 2929, 2929-31 (2000). Even assuming this study could be considered admissible evidence and that it has not been discounted in the past decade, it underscores my point. By looking at a booking photograph, the government could judge whether or not Pool is obese and, consequently, glean information regarding his risk of type-2 diabetes. And reference to Pool's gender shows his susceptibility to ovarian disorders is not an issue.
Not every disease has visible risk factors, of course, and genetic data is undoubtedly more precise than visual assessment in judging the probabilities associated with certain diseases. I remain unconvinced, however, that the information contained in the CODIS database at the present time categorically differs from the information already contained in the booking photo lineups on the bookshelf of virtually every police station in the country.
. I note with trepidation that the DNA Analysis Backlog Elimination Act does not facially limit DNA analysis to STRs or even non-coding DNA sequences. See 42 U.S.C. § 14135a(c)(2)(defining the DNA analysis allowed by the act as any “analysis of the deoxyribonucleic acid (DNA) identification information in a bodily sample”). Nonetheless, the executive branch has sensibly limited the information stored in CODIS, and it is that program we must evaluate today.
. There is no doubt that conviction is the key moment in reconfiguring the relationship between an individual’s privacy interests and the governmental interests that may impinge upon them. Nevertheless, we cannot blind ourselves to the various other gradations in our criminal justice system. Although United States v. Scott, 450 F.3d 863 (9th Cir.2006), clearly holds that a pre-trial releasee's "privacy and liberty interests [a]re far greater than a probationer’s,” id. at 873, Scott does not stand for the proposition that there is no meaningful distinction between an ordinary citizen and an individual for whom a judicial probable cause determination has been made. As this court acknowledged, "pretrial releasees must suffer certain burdens that ordinary citizens do not.” Id. at 872 n. 11. In this case, unlike Scott, we consider a restriction "designed to ensure [a defendant's] appearance in court.” Id.
We must not ignore admittedly finer distinctions in an effort to cram every individual into one of two categories — convict or citizen. Doing so will not remove the DNA profiling issue from the slippery slope; it would serve only to slicken the incline.