The defendant was charged with multiple offenses in connection with the deaths of his uncle, Thomas McNickles, and his cousin’s daughter, twelve year old Takeisha McNickles. He was convicted of murder in the second degree for the killing of Thomas,1 murder in the first degree (based on deliberate premeditation and felony-murder) for the killing of Takeisha, two indictments charging aggravated rape of Takeisha, and two indictments charging kidnaping.
Prior to trial, the defendant sought to exclude the results of deoxyribonucleic acid (DNA) testing, arguing that the testing did not satisfy the standards for validity and reliability set forth in Commonwealth v. Lanigan, 419 Mass. 15 (1994). After a seven-day evidentiary hearing, the judge, who was also the trial judge, ruled that the Commonwealth’s proffered DNA evidence met the requirements of Commonwealth v. Lanigan, supra, and allowed the Commonwealth to introduce that evidence at trial. On appeal, the defendant contends that the judge erred in admitting that DNA evidence. He also argues that one of the Commonwealth’s experts gave improper hearsay testimony concerning test results read by others at her laboratory. We conclude that the forensic evidence was properly admitted, and we affirm the defendant’s convictions.
1. Facts. On Friday afternoon, October 2, 1992, Takeisha’s mother took her to visit with her grandfather, Thomas, at his house on Devon Street in the Dorchester section of Boston. Over the course of that evening, various friends and family members came to Thomas’s house, including the defendant and Thomas’s brother, Marshall McNickles. By around 11 p.m., only Takeisha, Thomas, and the defendant were still there.
Shortly after midnight, a neighbor saw the defendant sitting in his automobile (a black Mercury Marquis) outside Thomas’s house. She invited him to go with her to a nearby crack house, *841which the defendant did. He stayed only briefly, however. When the neighbor returned home at around 1 a.m., she saw the defendant and Marshall standing in the doorway of Thomas’s house. They had a “dent puller,” a tool used in auto repair.2 She asked them, “What’s up?” They replied, “Nothing.”
The defendant’s girl friend and daughter lived with him in Norton. They were both home on the night of October 2, 1992. At some time during the night, the defendant’s daughter awoke to the sound of a “scream” from “a girl’s voice” coming from the back yard.3 Shortly thereafter, she heard two male voices from outside near the door to the cellar. She then saw the defendant’s black car, with two people in it, pull out of the driveway and drive off “fast” with the headlights turned off.
At around 6:30 a.m. on Saturday, October 3, 1992, Thomas’s body was discovered on Nightingale Street in Dorchester. Thomas had suffered multiple fatal blows to the head.4 From the condition of the body when found at 6:30 a.m., the medical examiner opined that the approximate time of death was sometime between 2 a.m. and 3 a.m.
The family was notified of Thomas’s death later that day. There was no sign of Takeisha nor any information about her. The defendant, in his first statement to the police that Saturday afternoon, claimed that Takeisha had asked Thomas to give her a ride home at around 2:15 a.m., that he had left Devon Street at that time with Thomas and Takeisha, that Thomas had dropped him off at his girl friend’s sister’s apartment on Hazelton Street so he could retrieve his car,5 and that Thomas and Takeisha had driven off in an unknown direction. The defendant claimed that he had then driven home to Norton.6
Late Saturday night, after the discovery of Thomas’s body *842and while Takeisha’s fate and whereabouts were still unknown, the defendant and Marshall went to see Mary Jones, another one of the defendant’s girl friends. The defendant told her of Thomas’s death, and told her that the killer had taken Thomas’s car and Takeisha. When asked whether Takeisha was also dead, the defendant replied that she was.7
That same night, the police found Thomas’s car a few blocks from his house. There were some blood stains on the rear passenger door, on the rear passenger quarter near the trunk, and on the pavement below the stained door. Takeisha was still missing.
The police spoke with the defendant again the next day, Sunday, October 4, 1992. He again claimed that his car had not been at Thomas’s on Friday night, and that he had gotten to Thomas’s house that night by borrowing a car from someone named “Larry” at work. He then corrected that version to state that “Larry” had dropped him off at Thomas’s on the night in question.8 He again claimed that he had left in Thomas’s car with Thomas and Takeisha, had been dropped off at Hazelton Street to pick up his own car, and had last seen Thomas drive off with Takeisha. The defendant acknowledged that he had not gone home to Norton. He claimed that he just “drove around” all night.
On the evening of October 4, the defendant telephoned Mary Jones and asked her to tell the police that he had been with her for the entire period of time from Thursday, October 1, through Saturday, October 3. Jones told the defendant that she “didn’t feel comfortable” giving a false alibi to the police. The defendant explained to Jones that he needed an alibi because, if *843he told the police where he had been, “it would be breaking his parole.”9
The next day, October 5, 1992, Takeisha’s body was finally found in a vacant, overgrown lot on Fabyan Street in Dorchester. Her top was pulled up; her pants were pulled down below her waist. She had been dead for some time, estimated to be at least one day. The cause of death was asphyxiation.
After the discovery of Takeisha’s body, the police questioned the defendant again. In this statement, the defendant admitted that he had taken his own car to Thomas’s the prior Friday night, but that he had then gone out, encountered his friend “Eddie” (whose last name he did not know), and loaned “Eddie” the car. “Eddie” dropped him off back at Thomas’s house, where he found Thomas asleep and Takeisha talking on the telephone. Takeisha was “giggling” and “flirting,” from which the defendant assumed she was talking with some boys. The defendant told Takeisha that she was “so fast” and that “[ejither me or my seventeen year old daughter should teach you.” He also described Takeisha as “physically mature for her age.” He again claimed that Takeisha had woken up Thomas and asked to be taken home, and that he had gone with Thomas and Takeisha as far as Hazelton Street to retrieve his own car. From there, however, he now claimed that he had driven over to and spent the rest of the night with Mary Jones.10
Based on his admission that he had been with Thomas and *844Takeisha just prior to the estimated time of Thomas’s death, and his shifting versions as to his whereabouts that night, the defendant was considered a suspect in these two homicides. In November, 1992, as the investigation continued, Thomas’s sister pleaded with the defendant to tell her anything he knew about how Thomas had died. The defendant told her, “I’m not taking this alone,” but gave no answer when she asked him what he meant. Another family member recalled the defendant saying that “he wasn’t going to do the time by himself.”
The police obtained blood samples from both the defendant and Marshall, and also retrieved samples from the bodies of Takeisha and Thomas. Those four known samples, plus vaginal swabs from Takeisha’s body,11 were sent to the Federal Bureau of Investigation (FBI) for DNA analysis. The FBI conducted restriction fragment length polymorphism (RFLP) testing on the samples, concluding that no definitive match could be made. The FBI did opine that the defendant could not be excluded as a contributor to the vaginal specimen taken from Takeisha’s body, and that both Thomas and Marshall were excluded as possible contributors to that specimen.
In May, 1993, shortly after receipt of the FBI’s report, the police also submitted specimens and samples to CBR Laboratories in Boston to conduct DNA testing by means of polymerase chain reaction (PCR). CBR was unable to obtain reliable results.12
As discussed in more detail below, the Commonwealth then submitted the samples to Cellmark Diagnostics (Cellmark), for further testing and analysis. Cellmark conducted PCR-based *845testing on the four known samples and on a rectal swab sample from Takeisha’s body, determined that a reliable reading could be made for one site, and opined that the defendant’s genotype at that site was the same as the contributor to the rectal swab sample. Cellmark also performed further analysis on the KFLP testing that had been conducted by the FBI. In Cellmark’s opinion, the results from three of the sites tested using RFLP were capable of more precise sizing and did match the defendant’s DNA. Based on the sizing performed by Cellmark, a statistician opined as to the likelihood that the defendant was a contributor to the sample retrieved from Takeisha’s vagina. Alleged errors in the Commonwealth’s presentation of these expert analyses are the sole issues now raised on appeal. For the reasons discussed below, we conclude that there was no error in the admission of this expert testimony.
2. DNA analysis using RFLP. The use of RFLP analysis of DNA has long been recognized as reliable, see Commonwealth v. Vao Sok, 425 Mass. 787, 799 (1997), and cases cited, and the defendant raises no challenge to the methodology itself.13 After the FBI interpreted the RFLP tests here as inconclusive, the Commonwealth asked Cellmark to review the FBI’s autoradiographs. Cellmark’s laboratory director, Dr. Robin Cotton, was able to size the DNA fragments (referred to as “bands”) on three of the four sites tested by the FBI. On the fourth site, known as D2S44, the quality of the autoradiograph did not permit accurate identification or sizing of all the bands.14 Dr. Cotton opined that the results at that particular site did not exclude the defendant (an opinion consistent with that of the FBI),15 but did not think that accurate sizing information could be reliably gleaned for that site due to the quality of the autoradiograph.
Working from the autoradiographs for the remaining three sites tested, Dr. Cotton concluded that the vaginal swab sample itself contained DNA from more than one person (referred to as *846a “mixed sample”). A swab sample taken from a rape victim will initially contain some cells from the victim herself. The process by which a laboratory attempts to separate the DNA of the victim from that of the contributor of any semen is not always successful in completely separating the two. As a result, the sample portion referred to as the “male fraction” may still contain some of the victim’s DNA. Thus, when DNA analysis reveals that DNA from more than one person is in the “male fraction” of a sample taken from a sexual assault victim, that result may be due to a small amount of the victim’s own DNA remaining in the sample following less than successful extraction techniques.
Here, the RFLP tests revealed some correlations between Takeisha’s DNA and the DNA in the mixed sample, such that she was viewed as a possible contributor.16 However, not all readings of Takeisha’s bands and the sample bands were clear, and Dr. Cotton could not say for certain that Takeisha was a contributor to the mixed sample.17
On the three sites for which Dr. Cotton was able to size the bands, the defendant remained as a possible contributor to the sample and both Marshall and Thomas were excluded as contributors. The actual size measurements determined by Dr. Cotton were provided to Dr. Bruce Weir, a population geneticist and statistician, for his analysis of the likelihood that the defendant was a contributor to that vaginal swab specimen. Because Dr. Cotton could only provide him with sizing information for three DNA sites, Dr. Weir’s calculation was based solely on the information from those three sites. In addition, because the evidence sample was itself a mixed sample, Dr. Weir’s *847analysis had to take into account the uncertainty as to the identity of the second contributor.
Dr. Weir analyzed the data using a statistical measure known as a “likelihood ratio.” A likelihood ratio compares the probability that the defendant was a contributor to the sample with the probability that he was not a contributor to the sample.18 Because it was unclear whether Takeisha’s own DNA was or was not in the evidence sample, Dr. Weir performed his calculations to cover both possibilities. Thus he calculated one likelihood ratio on the assumption that Takeisha’s DNA was not part of the sample, comparing the likelihood that the sample was comprised of the defendant’s DNA and another unknown person’s DNA with the likelihood that the sample was comprised of the DNA of two unknown persons. Assuming that Takeisha’s DNA was not part of the sample, it was 220 times more likely that the defendant was a contributor to the sample than that he was not a contributor. Then, Dr. Weir ran the calculations assuming that Takeisha was a contributor to the mixed sample, comparing the likelihood that the sample was comprised of the defendant’s DNA and Takeisha’s DNA with the likelihood that the sample was comprised of the DNA of an unknown person and Takeisha’s DNA. On that assumption, it was 37,000 times more likely that the defendant was a contributor to the sample than that another unknown person was the contributor.
The judge found that Dr. Weir’s application of likelihood ratios to mixed samples of DNA was “a valid and reliable scientific method of assigning statistical significance to DNA test results” on such samples. Her conclusion is subject to review for abuse of discretion. Canavan’s Case, 432 Mass. 304, 312 (2000). On appeal, the defendant contends that Dr. Weir’s analysis was unreliable in two respects. First, he contends that the statistical analysis should not have been calculated on *848information from only three sites when RFLP testing had been performed on four sites. Second, he argues that Dr. Weir’s calculation assuming that Takeisha’s DNA was in the sample was based on an assumption not established by the evidence. Neither argument convinces us that there was any abuse of discretion in introducing Dr. Weir’s calculations.
As to performing the analysis on only three instead of all four sites tested, the Commonwealth submitted extensive evidence explaining that the autoradiograph for the D2S44 site was not of sufficient quality to measure the bands with the requisite degree of accuracy. The autoradiograph for the D2S44 site did not exclude the defendant as a possible contributor, but it could not yield a more precise measurement of the bands at that site, and thus could not be compared with the population frequencies that Dr. Weir used to perform his statistical analysis. That the autoradiograph for one site was overexposed did not compromise the quality of the three remaining, and the lack of information about the D2S44 site did not prevent analysis of those sites for which there was reliable information. Where the defendant had not been excluded as a contributor by the results of testing at D2S44, there was nothing unreliable about analyzing the statistical significance of the matches that had been documented at three other sites.19
The defendant’s second argument about the expert’s reliance on unsupported assumptions is not persuasive as a matter of either logic or law. Precisely because the testing had suggested but not definitively confirmed the presence of Takeisha’s DNA in the mixed sample, Dr. Weir presented statistics for both possible scenarios. He did not “assume” anything, but rather presented statistical analyses for the full range of possibilities presented by the forensic evidence. Had the Commonwealth presented only a calculation based on the assumption that Takeisha’s DNA was part of the sample, the defendant’s argument would perhaps carry more weight. However, by presenting the exact same calculation covering the countervailing assump*849tian, the presentation was complete, and did not rely on any particular assumption or finding as to whether Takeisha’s DNA was or was not part of the sample.
Moreover, the evidence would support the conclusion that Takeisha’s own DNA was probably contained in the sample taken from her body. It is not surprising to lay persons, and it is confirmed by scientists, that a swab taken from a rape victim’s vagina will include some of the victim’s own DNA. Scientists acknowledge that their techniques for separating the victim’s DNA from the DNA in the perpetrator’s semen do not necessarily achieve complete separation. Where the resulting sample has the DNA of two persons, and where there appears to be some significant correlation between the victim’s DNA and certain bands in the mixed sample, a calculation that takes into account the potential that the victim’s DNA remains in the sample is an eminently reasonable calculation to present to the jury. The judge did not abuse her discretion in allowing the Commonwealth to present all of Dr. Weir’s analyses, including an analysis that took into account the strong likelihood that Takeisha’s own DNA was in the sample taken from her vagina.
3. Test of DQA1 site using PCR method. The Commonwealth also introduced results of another form of DNA analysis performed on rectal swab specimens from Takeisha. That analysis was performed using polymerase chain reaction (PCR) technology to amplify the sample and test for the specific alleles present at designated sites. At the Lanigan hearing (held prior to our decision in Commonwealth v. Vao Sok, supra), the defendant challenged the reliability of PCR methodology, and also contended that the results of this particular test were unreliable because the sample size was too small. With regard to the defendant’s first contention, thé reliability of PCR-based testing was upheld in Commonwealth v. Vao Sok, supra, and the judge correctly ruled that PCR testing satisfied the standards of Commonwealth v. Lanigan, supra.
The question whether there was sufficient sample size to yield reliable results from this particular PCR test was a separate issue. The defendant argues that, as to that issue, the judge abdicated her responsibility to determine reliability and erroneously held that the issue of sample size went only to the weight *850of the evidence, not to its admissibility. While we agree that the trial judge’s gatekeeper role under Commonwealth v. Lanigan, supra, includes the obligation to determine whether the testing at issue was conducted properly (and not just whether the testing method is theoretically reliable), the defendant’s argument about sample size did not, even if accepted, render these test results unreliable. Rather, as discussed below, the defendant conceded that the test results yielded some reliable, relevant information, challenging only whether the test could be read to yield a more precise result. As such, on the facts concerning this particular test, the differing opinions as to sample size did affect only the weight to be given these test results, not their admissibility, and the judge’s decision to admit the test results was not an abuse of discretion.
As explained in Commonwealth v. Vao Sok, supra at 791, PCR-based analysis is a method by which minute quantities of DNA may be replicated to create copies of DNA segments. The amplified DNA is then placed on test strips containing separate probes that turn color in the presence of specific alleles. A reading of the strip gives the genotype of the contributor by identifying both alleles at a designated site on the DNA.
The protocol for such testing requires that, in order to determine a genotype from the processed strip, the strip must “light up” a “control dot.” The control dot contains a probe for each of the possible alleles at that DNA site. Because there is a smaller amount of any given probe in the control dot than there is in any other dot, it is expected that the control dot will be the lightest of all the dots on the processed strip.
One of the possible reasons for the absence of any hue on the control dot is that the quantity of DNA tested is too small. While dots for specific alleles may show positive, the absence of a control dot leaves uncertainty as to whether other alleys were also present but were too minuscule in amount to be detected by their corresponding probes. Thus, as the defendant’s expert acknowledged, any dot that does fight up is an accurate indicator that that allele is present in the sample, but in the absence of a control dot, one cannot be certain that the test has established a base fine in order to detect all alleles present in the sample, and thus one cannot form an opinion as to the contributor’s correct genotype.
*851Here, Cellmark attempted PCR-based testing on the male fraction of the rectal swab sample at the DQA1 site and the five polymarker (PM) loci. There was no control dot showing on the PM strip for the rectal swab sample.20 Therefore, Cellmark did not interpret results at any of the PM loci.21
On the DQA1 strip, three Cellmark analysts had recorded that a faint control dot was visible on the wet strip. They therefore read the strip as yielding the genotype (i.e., as reflecting both alleles) of the contributor. The only probe that had activated was that for the allele known as 4.2/3, from which Cellmark concluded that the contributor was a homozygous 4.2/3 at the DQA1 site (i.e., that each of his two alleles was an identical 4.2/3). Tests of the defendant’s DNA confirmed that he is also a homozygous 4.2/3 at the DQA1 site. In the general population, one out of 1,000 Caucasians, one out of one hundred African-Americans, and one out of eighty-three Hispanics are homozygous 4.2/3 at the DQA1 site. Tests of the samples from Thomas and Marshall revealed that they were both 1.2/4.1 at the DQA1 site, thus excluding them as possible contributors.22 Because neither Thomas nor Marshall was a 4.2/3 at the DQA1 site, the test excluded Thomas and Marshall.
The defendant challenged whether the control dot on the DQA1 strip had activated and contended that the sample size on which PCR had been run was too small for reliable testing. *852Specifically, the defendant pointed out (and Dr. Cotton conceded) that there was no control dot visible on the dried strip or the photograph of the wet strip.23 In the opinion of the defense expert, the absence of the control dot in the photograph meant that the strip “is untypeable,” and could not be interpreted.
The defense expert also explained that the original sample size used was far smaller than that called for by any accepted protocol. The manufacturer of the kit used by Cellmark recommended a minimum of two nanograms DNA for its PCR-based tests at the PM sites, and scientific literature reported that one to two nanograms of DNA were needed for this kind of testing. The Commonwealth’s expert had participated in a study demonstrating that samples as small as .4 nanograms of DNA could be typed correctly at the DQA1 site using PCR-based testing. Here, however, the amount of DNA available for testing was unknown. The defense expert explained that small sample size could cause an allele to “drop[] out” during the PCR process, such that one allele would not register on the strip probe. Where the DQA1 strip from the rectal swab showed only one allele (4.2/3), and the photograph showed no control dot, the defense expert was of the view that it was not a “typeable result.”
However, the defense expert agreed that, notwithstanding small sample size and the absence of a control dot on the photograph, the strip reliably showed the presence of the 4.2/3 allele. He thus agreed that at least one of the contributor’s alleles at that site was a 4.2/3, and that the defendant therefore could not be excluded. What he disputed, however, was whether one could conclude that the contributor was homozygous. In his view, it was possible that the contributor had a second, different allele that had gone undetected. Thus, the strip provided some *853reliable information — that the contributor carried the 4.2/3 allele at the DQA1 site — but the defense expert could not confirm that the contributor was a homozygous 4.2/3 at the DQA1 site. This meant that the results at the DQA1 site were still consistent with the defendant (i.e., the contributor had at least one 4.2/3 allele at the DQA1 site and the defendant carries that same allele at that site), and the test results at that site also operated to exclude both Thomas and Marshall (as neither of them has a 4.2/3 allele at that site).
In order to admit the evidence of the DNA test results for the DQA1 site, the Commonwealth had to show not only that the underlying theory was scientifically valid, but also that the test itself was properly performed. See Commonwealth v. Vao Sok, 425 Mass. 787, 797-798 (1997), and cases cited. Here, Dr. Cotton agreed that, in the absence of a readable control dot, the probe strip should not be typed.24 However, in her view, a test was valid if the control dot was visible and, notwithstanding the small sample size, the presence of the control dot was determinative.25 Based on Cellmark’s assessment that a visible control dot made the test results on small samples sufficiently reliable, there was no abuse of discretion in allowing the Commonwealth to admit in evidence Cellmark’s determination of the DQA1 genotype of the contributor.
Moreover, the results of the PCR-based testing at the DQA1 site were, even accepting the defendant’s criticisms, of relevance to the case. Where, as here, the only readable PCR test results *854concerned a single site, the resulting information did not amount to or purport to be an actual identification of the defendant. Rather, the test yielded only descriptive information concerning the perpetrator. If the Commonwealth’s expert’s view were accepted by the jury, the descriptive information from the test was the perpetrator’s genotype at that one site (a homozygous 4.2/ 3), a trait shared by the defendant and many members of the general population.26 If the defense expert’s view of the test were accepted by the jury, the test still provided descriptive information about the perpetrator, i.e., that at least one of his alleles at the DQA1 site was a 4.2/3.27 That description, albeit not a complete genotype, was consistent with the defendant and inconsistent with both Thomas and Marshall.
That a scientific test provides descriptive information short of a definitive identification does not make it unreliable, irrelevant, or otherwise inadmissible.28 Very general descriptive information about a perpetrator (e.g., height, weight, hair color) is routinely admissible, even though the traits described are not unique to a defendant. The limited extent of PCR-based testing in this case yielded only such descriptive information, and was not presented to the jury as a complete identification or as a “match.”29 At no time did the Commonwealth exaggerate the significance of having the same type at the single DQA1 site or *855otherwise suggest that the PCR-based testing itself identified the defendant as the perpetrator.30 The amount of detail in the descriptive information from the DQA1 site varied based on which side’s expert the jury believed, but the test results themselves provided at least some reliable information about the perpetrator that was admittedly consistent with the defendant and that excluded two other possible suspects. We find no error in the judge’s decision to admit the results of PCR-based testing at the DQA1 site and to allow the jury to assess its weight.
4. Admission of hearsay evidence concerning the presence of the control dot. After prevailing at the Lanigan hearing, the Commonwealth called Dr. Cotton to testify at trial as to the results of the PCR-based testing at the DQA1 site. When the prosecutor asked Dr. Cotton her opinion as to the DQA1 genotype of the contributor to the male fraction of the rectal swab, the defendant interposed an objection, arguing that “the foundation needs to be established” (which he described as identifying “what it is she has reviewed, what was done, what she is relying upon”) before Dr. Cotton could render her opinion.31 Under Department of Youth Servs. v. A Juvenile, 398 Mass. 516, 531, 532 (1986), an expert is allowed to present an opinion based “on facts or data not in evidence if the facts or data are independently admissible” and are “of the sort that experts in that specialty rely on.” The facts concerning the dots on the wet strip, which had been observed by the analysts and noted on their scoring sheets, were potentially admissible, and there was nothing improper in the expert’s reliance on the observations that had been recorded by those analysts under *856Cellmark’s standard procedures at the time the tests were run. As such, the claimed lack of foundation for the expert’s opinion was not a proper basis for objection. See Commonwealth v. Sparks, 433 Mass. 654, 659 (2001).
The judge noted that Department of Youth Servs. v. A Juvenile, supra, would permit the expert to state her opinion, but defense counsel persisted in his objection about the need to provide a further “foundation.” In an apparent effort to avoid appellate issues, the prosecutor then offered to “establish” the “foundation” for the expert’s opinion. As part of that “foundation,” Dr. Cotton then explained the testing procedures, including the practice of having a third analyst review any strip where a dot was faint or hard to read. As such, the fact that the witness had not done the laboratory testing herself, and that she was relying on the observations of three analysts who reported having seen the control dot, was elicited during the expert’s direct examination.
The defendant now argues that the Commonwealth should not have been allowed to elicit from its expert this hearsay testimony about other analysts’ observation of a control dot. See Commonwealth v. Jaime, 433 Mass. 575, 577-578 (2001); Grant v. Lewis/Boyle, Inc., 408 Mass. 269, 273-274 (1990). The defendant’s argument correctly states the law, but ignores the fact that it was his own insistence that a “foundation” be laid that led to the error he now complains of. Under Commonwealth v. Jaime, supra, and Grant v. Lewis/Boyle, Inc., supra, the witness should have been allowed to testify to her opinion, but not to testify to the facts or data relied on in reaching that opinion when those facts or data were not yet themselves in evidence. Of course, the facts or data on which the expert has relied may then be elicited during cross-examination. Grant v. Lewis/Boyle, Inc., supra at 274. Department of Youth Servs. v. A Juvenile, supra at 532. Here, assuming that the Commonwealth had not presented this testimony to satisfy the defendant’s insistence that a “foundation” be laid, it is evident that the defendant intended to cross-examine the witness on this very subject, i.e., that she had not seen any control dot herself, that none was visible in the photograph, and that she was relying on the reports *857of others with respect to the alleged visibility of the control dot.32
We adhere to our position that an expert witness may not, on direct examination, present the specifics of hearsay information on which she has relied in reaching her opinion. Commonwealth v. Jaime, supra at 577. Grant v. Lewis/Boyle, Inc., supra. Here, however, where the defendant insisted that such specifics be presented on direct examination, and where the defendant was clearly intending to elicit those same specifics on cross-examination, there was no prejudicial error.33
5. Review under G. L. c. 278, § 33E. We have reviewed the entire record and decline to exercise our power under G. L. c. 278, § 33E, to order a new trial or direct the entry of a verdict of a lesser degree of guilt. Contrary to the defendant’s assertion, he was not convicted based on DNA evidence alone. He was, by eyewitness accounts and his own admission, with the victims immediately prior to their murders; his car was seen departing from his home shortly after a girl’s scream had been heard; he claimed knowledge of Takeisha’s death two days prior to her body’s being discovered; he gave multiple inconsistent statements to the police (and to family members) as to his means of transportation and whereabouts at the relevant time, statements that he now admits were false; he gave differing and implausible explanations for having given those false statements once their falsity was uncovered; and he made statements suggesting that he shared culpability for these crimes with another person. Two different methods of DNA testing then revealed that the defendant’s DNA was consistent with that of the person who had raped Takeisha. While the testing at the DQA1 site did not, by itself, provide enough information to conclude that the defendant was the perpetrator, the RFLP testing — long recognized as a valid and powerful method of comparing DNA samples — demonstrated a very high likeli*858hood that the seminal fluid found in the raped child’s body had come from the defendant.
Judgments affirmed.
Because the victims and various witnesses are related and share the same surname, we shall refer to them only by their first names.
Marshall had borrowed a dent puller a few days earlier from another family member who owned an auto body shop.
The defendant’s girl friend also recalled hearing a “human” noise, which she characterized as “chilling,” coming from the back yard.
The medical examiner opined that the injuries were consistent with being struck with a dent puller.
This explanation was contrary to the testimony of two witnesses who had seen the defendant’s car at Thomas’s house that night.
The defendant later admitted, and his girl friend and daughter confirmed, that he had not come home that night.
Later, when Jones confronted the defendant with the fact that he had known of Takeisha’s death prior to the discovery of her body, the defendant denied making such a statement.
The defendant’s boss at work, Lawrence Saunders, testified that he had neither loaned his car nor given a ride to the defendant that night. When the defendant was made aware that the police had been directed to Lawrence Saunders, the defendant told his girl friend that he had gotten a ride from another “Larry” who lived “next door” to Thomas. That neighbor, Larry Gethers, testified that he did not have an automobile at the time.
The jury was not informed that the defendant’s parole stemmed from a prior rape conviction, and they were instructed not to speculate about why the defendant was on parole.
When friends and family members asked the defendant why he had given these inconsistent versions to the police, he provided differing explanations for his admittedly false statements. To the girl friend with whom he lived, the defendant claimed that he had told these “stories” because he was worried they would “get in trouble” for using unregistered cars. However, all of the defendant’s shifting versions included his retrieving his car from Hazelton Street and driving somewhere with it, and nothing in his statements did anything to hide his use of that unregistered vehicle. On other occasions, he explained that he had been to a crack house and was reluctant to tell police because the involvement with drugs would violate his parole. However, the defendant’s brief trip to a crack house had concluded between midnight and 1 a.m., prior to his departure with the victims. An accurate accounting of his whereabouts and activities from 2 a.m. onward would not have implicated him in illegal drug activity that had concluded prior to that time.
Vaginal swabs had tested positive for the presence of semen, confirming the suspicion that Takeisha had been raped.
The tests of CBR Laboratories were excluded from admission as part of the consolidated review of the reliability of PCR-based testing for this case and three other pending cases. After another judge ruled that PCR-based testing at the polymarker (PM) sites was not reliable, the Commonwealth appealed from that ruling in the other cases, but did not appeal from that ruling with respect to this defendant. On appeal in the other cases, this court held that PCR-based testing at the PM sites was scientifically valid and reliable. Commonwealth v. Vao Sok, 425 Mass. 787 (1997). We noted there that the Commonwealth had not taken an appeal with respect to this defendant, and that the Commonwealth was conducting further testing in the present case. Id. at 788 n.2.
For a detailed description of the theory and process of KFLP analysis, see the appendix to Commonwealth v. Curnin, 409 Mass. 218, 227-231 (1991).
The autoradiograph for D2S44 had been overexposed during the FBI’s testing.
The defendant’s experts agreed that the D2S44 autoradiograph could not be read as a definitive exclusion of the defendant.
The defendant’s expert acknowledged that Takeisha was a possible contributor to the sample.
The analysis of the vaginal swab sample and the known sample from Takeisha was complicated by the fact that her body had gone undiscovered for a considerable period of time. The experts referred to the DNA in these samples as “degraded,” meaning that the DNA had begun to break down from natural processes prior to the samples having been retrieved. While RFLP analysis can be applied to degraded DNA material, that degradation complicates the interpretation of the resulting autoradiographs. Instead of clear bands on the known sample from Takeisha, the autoradiographs were somewhat smeared due to sample degradation, making the identification of her bands less certain.
The utility of such an approach (and in particular its suitability for analysis of mixed samples) has long been recognized. See National Research Council, The Evaluation of Forensic DNA Evidence 129 (1996) (“when the contributors to a mixture are not known or cannot otherwise be distinguished, a likelihood ratio approach offers a clear advantage and is particularly suitable”). The defendant does not contend that the use of likelihood ratios is an unreliable method for analyzing DNA results.
Dr. Weir noted that a match at three different sites “happens very rarely.” The defendant’s expert did not contend that three sites were an insufficient basis for analysis, but expressed only the view that any analysis should consider all sites tested because “you have to use all the evidence.”
Cellmark has two analysts check each strip independently, and if those two analysts do not agree or are unsure about a reading, a third analyst reviews the strip. Three analysts read the PM strip on the male fraction of the rectal swab sample. Two analysts reported that they could not detect any change in hue at the control dot. One analyst reported that color was visible on the control dot. Where two out of the three analysts detected no visible ' color on the control dot, Cellmark treated the strip as showing no control dot.
In the absence of a control dot, the presence of an inconsistent allele can still be used to exclude a suspect, even though the strip cannot be used to determine the actual genotype of the specimen contributor. Here, Cellmark did confirm that there was no inconsistency at any of the PM sites and that the results therefore did not exclude the defendant. However, in the absence of the control dot, Cellmark would not opine as to the results, and the jury were not presented with any comparison of the defendants PM types with the PM test results from the male fraction of the rectal swab sample.
Takeisha herself was a homozygous 1.2 at DQA1, and thus was not a contributor to this sample. Although the defendant questioned whether the sample actually contained semen, the presence of another person’s DNA in the rectal swab sample confirmed that Takeisha had also been anally raped.
Dr. Cotton explained that the reading is done by analysts when the strip is wet, at which time the color on each probe will be at its darkest; that the photograph is taken while the strip is still wet; that the strip fades as it dries; and that faint colors visible to the naked eye are not always visible on a photograph. Thus, notwithstanding the absence of any visible control dot on the photograph, Dr. Cotton was willing to rely on the observations made by three analysts reading the wet strip, each of whom had recorded seeing a control dot.
For that very reason, the Commonwealth did not seek to introduce the Cellmark test results from the PM loci, as no control dot appeared on the PM strip. See note 21, supra.
The defendant argues that Dr. Cotton merely opined that she was “comfortable” using PCR testing on a small sample size, and that her conclusory assertion about feeling “comfortable” is not sufficient to establish scientific reliability. The use of the term “comfortable” came from the defendant’s own cross-examination of Dr. Cotton. (We note that the defendant’s expert also used the term “comfortable” to describe how he felt about testing of various sample sizes.) That some nonscientific terminology occasionally creeps into the questioning of an expert does not undermine the reliability of the scientific process being discussed by the expert. Dr. Cotton herself explained why, in her view, the presence of the control dot was a sufficient indicator of the test’s reliability, even for very small samples. Her discussion of the issue was not limited to a claim that she was “comfortable” with this particular sample size.
The frequency of a homozygous 4.2/3 at the DQA1 site varies from a low of one in 1,000 in the Caucasian population, to one in 100 in the African-American population, to as high as one in eighty-three in the Hispanic population. Defense counsel’s cross-examination of Dr. Cotton walked through the arithmetic of how many out of 1,000 African-Americans, 10,000 African-Americans, 100,000 African-Americans, and one million African-Americans would have this same genotype, and then pointedly asked her if she knew how many African-Americans lived in Boston.
There are seven alleles at the DQA1 site, resulting in twenty-eight possible genotypes. Seven of those possible genotypes incl'tide at least one 4.2/3 allele. The defendant could have, but did not, introduce information indicating what percentage of various population subgroups carry at least one 4.2/3 allele at the DQA1 locus.
For example, although blood typing provides results that are far less specific than a DNA match, serologists routinely testify concerning blood typing of crime scene evidence that is consistent with a defendant’s blood type. See, e.g., Commonwealth v. Sparks, 433 Mass. 64, 660 (2001).
Nor were the PCR results included in any of Dr. Weir’s statistical calculations. Those calculations, which did result in strong statistical evidence against the defendant, were based exclusively on RFLP testing.
Indeed, at the very beginning of her closing argument, the prosecutor told the jury, “I’d like to just remind you that DNA did not identify [the defendant] as a suspect in this case.” She then reviewed the many suspicious circumstances, including the defendant’s multiple inconsistent statements, that pointed to the defendant’s involvement in these crimes. In her later summation of the DNA evidence, the prosecutor emphasized to the jury that RFLP testing was “more discriminating” and “more sensitive.” When referencing the PCR-based results at DQA1, the prosecutor’s closing reminded the jury that one per cent of the African-American population had the defendant’s genotype at that site.
From the detailed presentation at the earlier Lanigan hearing, defense counsel knew that Dr. Cotton was relying on scoring sheets from three analysts who indicated that they had seen the control dot when the DQA1 strip was wet.
Defense counsel did cross-examine Dr. Cotton extensively on these points.
As discussed above, the presence or absence of the control dot affected only the assessment of the contributor’s genotype, not the determination of the presence of the 4.2/3 allele in the sample. That the dispute about the visibility of the control dot surfaced during the expert’s direct examination, instead of awaiting cross-examination, would have no impact on the jury’s assessment of the weight to be accorded this particular DNA test.