660 So.2d 257 (1995)
Robert HAYES, Appellant,
v.
STATE of Florida, Appellee.
No. 79997.
Supreme Court of Florida.
June 22, 1995.
Rehearing Denied September 13, 1995.
Richard L. Jorandby, Public Defender and Richard B. Greene, Asst. Public Defender, West Palm Beach, for appellant.
Robert A. Butterworth, Atty. Gen. and Sara D. Baggett, Asst. Atty. Gen., West Palm Beach, for appellee.
OVERTON, Justice.
This is an appeal by Robert Hayes from a conviction of first-degree murder and a sentence of death imposed by the trial judge in accordance with the jury's recommendation. We have jurisdiction. Art. V, § 3(b)(1), Fla. Const. While the record contains evidence suggesting that Hayes committed the homicide in this case, it also contains objective physical evidence suggesting that someone other than Hayes was responsible. We find that the errors committed during Hayes' trial were not harmless and, for the reasons detailed in this opinion, we conclude that Hayes' conviction of first-degree murder must be reversed, his death sentence vacated, and this case remanded for a new trial. In this opinion, this Court addresses for the first time how deoxyribonucleic acid (DNA) test results may be admitted in the trial courts of this State.
The record reveals that, on the morning of February 20, 1990, the victim, a female groom at the Pompano Harness Track, failed to show up for work. Security guards at the track went to the victim's room in a dormitory near the track where the victim and other female grooms lived. After opening the locked door, the security guards found the victim's body on the floor of her room. She was wearing only blue jeans and a T-shirt. A tank top shirt was found lying on the floor. An investigation of the homicide led to the arrest of Robert Hayes, an African-American groom who also worked at the harness track.
At Hayes' trial, a medical examiner testified that the victim was killed by manual strangulation and explained that she had bruises on both sides of her head and her upper lip. He also testified that he found no bruising to the vaginal or anal area and that the victim had light brown hair clutched in one of her hands. The State presented the testimony of a serologist concerning tests conducted on a vaginal swab taken from the victim, as well as on samples cut from the tank top found on the floor of the victim's room. The serologist's testimony revealed that seminal fluid was present on both the tank top and the vaginal swab. However, no seminal fluid was found on the victim's blue jeans.
A technician from Life Codes, a DNA testing company, testified that she performed DNA testing on samples taken from the victim, the defendant, the vaginal swab, and the tank top. She explained that in doing her testing she worked without anyone watching her and that she also ran tests on another case at the same time. She stated that testing on the vaginal swab produced a seven-band DNA match with the blood sample taken from the defendant and that testing on the tank top sample produced a three-band match. On cross-examination, counsel for the appellant challenged the testing methods used by the technician. Defense counsel also presented expert testimony challenging the DNA test results.
The prosecutor also presented the testimony of several of the victim's co-workers. One of the co-workers testified that she saw a man, whom she later identified as Hayes, at the door of the victim's room at about 8:45 on the evening of the murder. She testified that she heard the victim telling Hayes that she was going into her room to go to sleep and that [Hayes] was not going to come in. Another co-worker testified as follows:
Q [Prosecutor] The police came to you? The detectives came to you and asked you ultimately what you had heard?
A No, no. What happened was somebody told me that works for the track that [the victim] had been murdered and I called John Beatrice, which is security, and told him I know who did it.
Q Okay.
....
Q You gave Bob, a name of the defendant; is that correct?
A Yes.
Defense counsel preserved the issue for review through a proper objection. It was later determined that this witness had no direct knowledge that Hayes was involved in the crime.
The State also presented the testimony of another female groom who stated that she was attacked by Hayes approximately eighteen months prior to this offense at a racetrack in New Jersey. She testified that she and Hayes had gone out for dinner, then to a bar for drinks, and then to her dormitory room where they talked. She stated that, while they were so engaged, Hayes attacked her, got her down on the floor on her stomach, and proceeded to choke her. She finally got him to release her and allow her to go to the rest room down the hall. She called a security officer and Hayes was arrested for simple assault. The charge was later dropped. No evidence was presented at trial that this was a sexual assault.
Another prosecution witness testified concerning an altercation between Hayes and the victim one month before the homicide. The prosecution next presented various statements Hayes made after his arrest. Hayes had told the police immediately after his arrest that he last saw the victim around 5:30 on the evening of the murder. Later, after being confronted with the statement of an eyewitness who had seen Hayes with the victim later in the evening, Hayes changed his story and admitted speaking to the victim after 5:30 p.m.; however, he denied ever going into the victim's room. The prosecution also presented a jailhouse informant who testified that Hayes admitted to him the raping and strangling of the victim.
Hayes presented evidence by several witnesses. The lead investigator in the case testified that, although the police had found hair samples and fingerprints at the crime scene, none matched Hayes'. A hair analyst from the sheriff's office testified that the hair found clutched in the victim's hand was inconsistent with Hayes' hair. Hayes also presented expert testimony that substantially challenged the procedures used by the company that conducted the DNA tests. Finally, Hayes presented the testimony of another female groom who stated that she had spoken with the victim at ten o'clock on the evening of the murder, approximately an hour after Hayes was last seen speaking with the victim, and that the victim was not distressed in any way.
The jury found Hayes guilty of first-degree murder and the judge sentenced him to death. Hayes then brought this appeal in which he raises multiple claims of error. We find that the following four issues merit discussion: (1) whether error was committed in admitting collateral crime evidence; (2) whether the DNA test results were unreliable because of the laboratory procedures employed; (3) whether a prosecution witness was erroneously allowed to present hearsay evidence of a statement of the victim's and to give her opinion that Hayes committed this offense; and (4) whether the prosecution was improperly allowed, over objection, to place the burden on Hayes to do certain scientific testing.
Collateral Crime Evidence
As previously stated, the prosecution presented collateral crime evidence that Hayes had previously attacked another female groom who was working with Hayes at a racetrack in New Jersey. This incident occurred eighteen months prior to the murder in this case. The victim of this prior offense testified that, after she and Hayes had gone out for drinks and dinner, and while they were talking in her dormitory room, he suddenly jumped on her, held her on the ground on her stomach, choked her, and then released her. The witness stated that Hayes then allowed her to sit up and leave the room to go to the rest room down the hall. Although Hayes was arrested for simple assault, the charge was later dropped. In this case, Hayes sought, by a motion in limine, to exclude this evidence. The trial judge denied the motion but subsequently recognized a continuing objection by Hayes to preserve the issue for appellate review.
The Evidence Code, under section 90.404(2)(a), Florida Statutes (1993), allows a party to introduce similar fact evidence of other crimes when it is relevant to prove a material fact in issue. In Drake v. State, 400 So.2d 1217 (Fla. 1981), we set forth the principles of how this evidentiary provision should be applied. See also Thompson v. State, 494 So.2d 203 (Fla. 1986); Peek v. State, 488 So.2d 52 (Fla. 1986). In Drake, we stated:
Williams v. State [110 So.2d 654 (Fla.), cert. denied, 361 U.S. 847, 80 S.Ct. 102, 4 L.Ed.2d 86 (1959)] holds that evidence of similar facts is admissible for any purpose if relevant to any material issue, other than propensity or bad character, even though such evidence points to the commission of another crime. The material issue to be resolved by the similar facts evidence in the present case is identity, which the State sought to prove by showing Drake's mode of operating.
The mode of operating theory of proving identity is based on both the similarity of and the unusual nature of the factual situations being compared. A mere general similarity will not render the similar facts legally relevant to show identity. There must be identifiable points of similarity which pervade the compared factual situations. Given sufficient similarity, in order for the similar facts to be relevant the points of similarity must have some special character or be so unusual as to point to the defendant.
Drake, 400 So.2d at 1219 (emphasis added).
In the instant case, the State sought to prove the identity of the murderer by showing a pattern of allegedly similar behavior by Hayes on a prior occasion. We conclude that, consistent with Drake, there are insufficient points of similarity to the instant offense to warrant admitting evidence of the previous attack. We note that the victim in the prior offense had voluntarily gone out with Hayes before she and Hayes returned to her room, that the victim did not testify that Hayes had sexually assaulted her, that Hayes was charged with only a simple assault (a charge that was later dropped), and that Hayes released the victim and allowed her to leave the room. We also find that any marginal relevance the prior attack may have had to the instant case was substantially outweighed by its prejudicial effect. See § 90.403, Fla. Stat. (1993). In fact, the differences in this case are considerably greater than the differences in Drake. Accordingly, we find that the trial judge erroneously admitted the evidence of the previous attack.
Admissibility of DNA Evidence
In his second claim of error, Hayes challenges the admissibility of the results of the two DNA tests allowed into evidence in this case. This claim presents an issue of first impression with this Court. While other Florida courts have recognized the admissibility of DNA evidence as a new but proper scientific procedure, see, e.g., Brim v. State, 654 So.2d 184 (Fla. 2d DCA 1995); Vargas v. State, 640 So.2d 1139 (Fla. 1st DCA 1994), review granted, No. 83,935 (Fla. May 16, 1995)[1]; Andrews v. State, 533 So.2d 841 (Fla. 5th DCA 1988), review denied, 542 So.2d 1332 (Fla. 1989), this Court has not yet addressed the issue. We have recently made it clear that Florida utilizes the Frye[2] test to determine the admissibility of new or novel scientific evidence such as DNA. Ramirez v. State, 651 So.2d 1164 (Fla. 1995); see also Flanagan v. State, 625 So.2d 827 (Fla. 1993); Stokes v. State, 548 So.2d 188 (Fla. 1989). We stated in Ramirez that the admission into evidence of expert opinion testimony of a new scientific principle requires a four-step inquiry. Ramirez, 651 So.2d at 1166. The trial judge must determine whether: (1) expert testimony will assist the jury in understanding the evidence or in determining a fact in issue; (2) the expert's testimony is based on a scientific principle or discovery that is "sufficiently established to have gained general acceptance in the particular field in which it belongs" under the Frye test; and (3) the particular expert witness is qualified to present opinion evidence on the subject in issue. If the answer to the first three questions is in the affirmative, the trial judge may proceed to step four and allow the expert to present an opinion to the jury. Id. The jury then determines the credibility of the expert's opinion, which it is free to either accept or reject. Id.
The admissibility of DNA evidence in the courts throughout the country has been an issue of considerable interest and concern. Because of the substantial questions surrounding DNA typing, reliability, and methodology standards, the National Research Council of the National Academy of Sciences was called upon, in 1989, to establish recommended standards and methodology concerning DNA testing. Victor A. McKusick, Preface to Committee on DNA Technology in Forensic Sciences, National Academy of Sciences, DNA Technology in Forensic Science at vii (1992). The National Research Council published its report and recommendations in 1992. See id. This report makes it clear that courts should take judicial notice of three scientific underpinnings of DNA typing.[3] The National Research Council emphasized the importance of the scientific testing methods used in DNA typing and stated:
Forensic DNA analysis should be governed by the highest standards of scientific rigor in analysis and interpretation. Such high standards are appropriate for two reasons: the probative power of DNA typing can be so great that it can outweigh all other evidence in a trial; and the procedures for DNA typing are complex, and judges and juries cannot properly weigh and evaluate conclusions based on different standards of rigor.
Id. at 51-52. This report emphasizes the importance of the testing protocol used in DNA analysis and has a separate section that explains in detail how DNA evidence is appropriately admissible under the Frye test. It emphasizes the application of the Frye test to the testing procedures utilized in this complex process and states that "a court should recognize that the expertise of more than one discipline might be necessary to explain [the procedures.]" Id. at 133 (Emphasis added.) The report explains that, in applying the Frye test to DNA testing procedures, there are four pertinent assumptions:
Assumption 1 that, with the exception of identical twins, each person's DNA is unique is so well established in human molecular genetics that a court is justified in judicially noticing it, even in the context of a Frye hearing.
Assumption 2 concerns the validity of procedures for extracting DNA from samples of blood, semen, and other materials and analyzing it for the presence and size of polymorphisms. With regard to application in scientific research, the validity is sufficiently well established in the case of RFLP analysis with Southern blots that judicial notice is also appropriate. With regard to the application in forensic science, however, additional questions of reliability are raised. For example, forensic DNA analysis frequently involves the use of small, possibly contaminated samples of unknown origin, such as a dried blood stain on a piece of clothing. Some experts have questioned the reliability of DNA analysis of samples subjected to "crime scene" conditions. In addition (as noted in Chapters 2 and 3), the details of the particular techniques used to perform DNA typing and to resolve ambiguities evoke a host of methodological questions. It is usually appropriate to evaluate these matters case by case in accordance with the standards and cautions contained in earlier portions of this report, rather than generally excluding DNA evidence. Of particular importance once such a system of quality assurance is established would be a demonstration that the involved laboratory is appropriately accredited and its personnel certified. Some aspects (such as the validity of the theory underlying RFLP analysis) might be so well established that judicial notice is warranted. Others (such as quantitative correction of band shifting with a single monomorphic fragment) might not be sufficiently well established to justify admissibility.
Assumption 3 related to the adequacy of statistical databanks used to calculate match probabilities rests on unproven foundations. Many experts question the adequacy of current databanks for making probability estimates, and the use of multiplicative modes of combining probabilities are also open to serious question (see Chapter 3). The solution, however, is not to bar DNA evidence, but to ensure that estimates of the probability that a match between a person's DNA and evidence DNA could occur by chance are appropriately conservative (as described in Chapter 3).
The validity of assumption 4 that the analytical work done for a particular trial comports with proper procedure can be resolved only case by case and is always open to question, even if the general reliability of DNA typing is fully accepted in the scientific community. The DNA evidence should not be admissible if the proper procedures were not followed. Moreover, even if a court finds DNA evidence admissible because proper procedures were followed, the probative force of the evidence will depend on the quality of the laboratory work. More control can be exercised by the court in deciding whether the general practices in the laboratory or the theories that a laboratory uses accord with acceptable scientific standards. Even if the general scientific principles and techniques are accepted by experts in the field, the same experts could testify that the work done in a particular case was so flawed that the court should decide that, under Frye, the jury should not hear the evidence.
Id. at 133-34 (emphasis added).
In the instant case, the DNA test on the tank top was inconclusive until the Life Codes technician applied the controversial "band-shifting" technique. The technician stated that the samples taken from the tank top produced a three-band match with the samples taken from the defendant. The defense challenged this conclusion, not only on cross-examination but also by presenting an expert who testified that a three-band match was not truly a "match," and that corrections made due to band-shifting were not accepted in the scientific community. The National Research Council, in its report, agrees with the expert for the defendant and directly addresses the problem of band-shifting as follows:
Testing for band shifting is easy, but correcting it is harder....
....
Little has been published on the nature of band shifting, on the number of monomorphic internal control bands needed for reliable correction, and on the accuracy and reproducibility of measurements made with such correction. For the present, several laboratories have decided against attempting quantitative corrections; samples that lie outside the match criterion because of apparent band shifting are declared to be "inconclusive." The committee urges further study of the problems associated with band shifting. Until testing laboratories have published adequate studies on the accuracy and reliability of such corrections, we recommend that they adopt the policy of declaring samples that show apparent band shifting to be "inconclusive."
Id. at 60-61 (emphasis added). When a major voice in the scientific community, such as the National Research Council, recommends that corrections made due to band-shifting be declared "inconclusive," we must conclude that the test on the tank top is unreliable. Our holding in this regard is not without precedent. In People v. Keene, 156 Misc.2d 108, 591 N.Y.S.2d 733, 740 (N.Y. Sup. Ct. 1992), that court relied in part on the National Research Council report to exclude DNA test results that were tainted by band shifting. It is clear from the record that the methodology used by the technician in this case was not sufficiently established to have gained general acceptance in the scientific community under the Frye test. Accordingly, we must hold that the DNA match concerning the tank top was inadmissible as a matter of law.
The seven-band DNA match on the sample taken from the vaginal swab presents a different question. It was also challenged by Hayes on cross-examination and through his expert who testified that the procedures used by Life Codes did not protect against the risk of contamination. We do not find that this DNA evidence should be excluded as a matter of law, but we also do not find that we should approve it for admission under the circumstances of the record in this case because we find that the Frye test was not properly applied, particularly as suggested in the National Research Council report. We do not fault the trial judge in this instance who did not have the benefit of the National Research Council report. DNA test results as evidence in criminal trials are not only new, but, as important, such results are based on technology that is still evolving and must be evaluated on a case by case basis. This evolving technology is constantly changing as evidence by the fact that the National Research Council is presently revising its 1992 report.[4] Without question, DNA testing methodology, while an extremely important new identification technique, has not yet reached the level of stability of other forms of identification such as fingerprint comparisons. In the retrial of this defendant, the DNA evidence pertaining to the vaginal swab may be presented if the State can establish that the methodology utilized by the technician in performing the test meets the requirements of the Frye test.
In summary, we find that the DNA evidence would assist the jury in this case in determining a fact in issue. We take judicial notice that DNA test results are generally accepted as reliable in the scientific community, provided that the laboratory has followed accepted testing procedures that meet the Frye test to protect against false readings and contamination. With regard to the tank top, we find that the DNA test was inadmissible because it did not meet accepted scientific principles. Finally, we find that, while this record does not support a proper application of the required Frye test to the procedures utilized to obtain the DNA test results on the vaginal swab, the DNA evidence may be presented upon retrial subject to a proper finding under the Frye test.
Remaining Issues
In his third claim of error, Hayes asserts that the trial judge should have sustained a defense objection when a witness for the State testified that she knew the defendant had murdered the victim because the victim had told her that she feared the defendant and his threats. These statement were not based on personal knowledge, were unresponsive to the prosecutor's question, and included both inadmissible hearsay and an opinion on the ultimate fact. The objection made by the defense to this testimony was sufficient to preserve this error for review. We find that the trial judge erred in overruling the defense objection.
The fourth error occurred when the trial court allowed the prosecutor to elicit testimony, over defense objection, concerning the failure of the defense to request testing of various pieces of scientific evidence. The State called as a witness an employee of the Broward County Sheriff's Office crime lab who testified concerning various pieces of physical evidence found at the scene of the murder, including clothing stained with blood. On cross-examination, the defense brought out the fact that the State had never requested a test of the blood stains. The apparent goal of this line of questioning was to cast doubt on the thoroughness of the State's investigation and to imply that a test of the blood could have eliminated Hayes as a suspect. Then, on redirect, the trial judge overruled a defense objection and allowed the State to inquire whether the defense had requested any testing of the blood stains. The witness replied that the defense had not asked the crime lab to test the blood stains and added that the lab had complied with such requests in the past for other defense attorneys. Similar comments were made by the prosecutor in closing argument concerning the failure of the defense to test hairs found at the scene of the murder. We find that the prosecutor's comments were prejudicial. In Jackson v. State, 575 So.2d 181, 188 (Fla. 1991), we found that the State had erred when the prosecutor commented on the defendant's failure to call a particular witness to testify. We explained:
It is well settled that due process requires the state to prove every element of a crime beyond a reasonable doubt, and that a defendant has no obligation to present witnesses. Accordingly, the state cannot comment on a defendant's failure to produce evidence to refute an element of the crime, because doing so could erroneously lead the jury to believe that the defendant carried the burden of introducing evidence.
Id. The prosecutor's questions and statements in the instant case may have led the jury to believe that Hayes had an obligation to test the evidence found at the scene of the murder and to prove that the hair and blood samples did not match his own. Clearly, Hayes had no such obligation.
The State asserts that its comments were invited by the defense in opening arguments when counsel told the jury that it would receive probative evidence that the defendant did not commit the crime. In Jackson, we recognized an exception to the general rule that the prosecutor may not comment on the failure of the defense to call any witnesses when the defendant raises an issue for which the defense carries some burden of proof. Jackson, 575 So.2d at 188. For example, if a defendant claims an alibi, we have held that the State may comment on the failure of the defendant to call witnesses to substantiate the alibi. Buckrem v. State, 355 So.2d 111, 112 (Fla. 1978). However, the exception cited by the State does not apply in the instant case because Hayes never put at issue any claim for which he carried any burden of proof and for which the prosecutor's comments would be relevant. In opening argument defense counsel stated: "You will be provided with solid physical evidence that this terrible crime was committed by another person. In fact, you will receive evidence that the murderer was not a black man but a white person." The evidence to which defense counsel referred was the hair found clutched in the victim's hand. These strands were inconsistent with the defendant's hair and suggested that a Caucasian individual may have committed the murder. Defense counsel never assumed any responsibility for presenting the hair strands to the jury as part of an affirmative defense. In fact, it was the prosecutor who first raised the issue of the hair strands in opening argument, and defense counsel's statement is consistent with the notion that evidence presented by the State itself would be probative evidence that another person committed the murder.
We also find no merit to the State's contention that any error in regard to this issue was remedied when the judge gave a curative instruction to the jury. The curative instruction informed the jury that, although the defense had no obligation to test the evidence collected at the crime scene, it did have the opportunity to have it tested. "Opportunity" in this context implies an obligation and we are unwilling to assume that the jury could have found a measurable distinction between the terms.
While evidence exists in this case to establish that Hayes committed this offense, physical evidence also exists to establish that someone other than Hayes committed the offense. The prosecution focused its final argument, in a large part, on the evidence that we have found was erroneously admitted. We conclude that the errors we have identified cannot collectively be found to be harmless beyond a reasonable doubt. Accordingly, we reverse Hayes' conviction, vacate his death sentence, and remand for a new trial. Hayes' remaining issues are either without merit[5] or mooted by our disposition of this appeal.[6]
It is so ordered.
GRIMES, C.J., and SHAW, KOGAN, HARDING, WELLS and ANSTEAD, JJ., concur.
NOTES
[1] Brim and Vargas each deal with the admissibility of evidence concerning the statistical likelihood that someone other than the defendant has a DNA pattern that matches the DNA taken from the crime scene. This particular aspect of the admissibility of DNA evidence is not at issue in the instant case.
[2] Frye v. United States, 293 F. 1013, 1014 (D.C. Cir.1923).
[3] The report states:
Courts should take judicial notice of three scientific underpinnings of DNA typing:
(1) The study of DNA polymorphisms can, in principle, provide a reliable method for comparing samples.
(2) Each person's DNA is unique (except that of identical twins), although the actual discriminatory power of any particular DNA test will depend on the sites of DNA variation examined.
(3) The current laboratory procedure for detecting DNA variation (specifically, single-locus probes analyzed on Southern blots without evidence of band shifting) is fundamentally sound, although the validity of any particular implementation of the basic procedure will depend on proper characterization of the reproducibility of the system (e.g., measurement variation) and inclusion of all necessary scientific controls.
[4] See Federal Judicial Center, Reference Manual on Scientific Evidence 309 (1994); Rorie Sherman, New Scrutiny for DNA Testing, Nat'l L.J., Oct. 18, 1993, at 3.
[5] Hayes has raised the following issues which we find to be without merit: (1) whether the trial court erred in failing to grant a judgment of acquittal as to premeditated murder; and (2) whether Florida's death penalty statute is unconstitutional.
[6] Hayes has raised the following issues which are mooted by our disposition of this appeal: (1) whether the trial court erred in overruling Hayes' objection to the prosecutor's inflammatory closing argument; (2) whether the trial court erred in denying Hayes' motion for a new trial based on newly discovered scientific evidence; (3) whether the trial court erred in admitting collateral offense evidence as the charge had been dismissed; (4) whether the trial court erred in restricting the cross-examination of a key prosecution witness; (5) whether the trial court erred in restricting questioning of the jurors on the subject of racial prejudice; (6) whether the trial court erred in allowing the prosecution to excuse a black juror over objection; (7) whether the trial court erred in denying Hayes' motion to suppress his statements to police; (8) whether the trial court erred in giving undue weight to the jury's death recommendation; (9) whether the trial court erroneously applied a presumption of death; (10) whether the trial court erred in failing to consider and find the non-statutory mitigating factor that this was an offense with little or no premeditation; (11) whether the trial court erred in finding that the capital felony was especially heinous, atrocious, or cruel; (12) whether section 921.141(5)(d), Statutes (1989), is unconstitutional on its face and as applied in this case; and (13) whether death is disproportionate.