Evaluation of nucleosome forming potentials (NFPs) of forensically important STRs

Article Outline

• Abstract
• 1. Introduction
• 2. Materials and methods
  • 2.1. Selecting markers and obtaining base sequences
  • 2.2. NXSensor mechanisms and parameters
  • 2.3. NuScore mechanisms and parameters
  • 2.4. Comparisons of original base sequences with random arrangements
• 3. Results
  • 3.1. NXSensor
  • 3.2. NuScore
  • 3.3. Comparisons of original base sequences with random arrangements
• 4. Discussion
• 5. Conclusion
• References
• Copyright

Abstract 

Degraded forensic samples have proved difficult to analyze and interpret. New analysis techniques are constantly being discovered and improved but researchers have overlooked the structural properties that could prevent or slow the process of degradation. In theory, DNA that are bound to histones as nucleosomes are less prone to degradation, because nucleosomes prevent DNA from being exposed to degradative enzymes. In this study we determined the probability of 60 forensic DNA markers to be bound to histones based on their base sequence composition. Two web-based tools – NXSensor and nuScore – were used to analyze four hundred base pairs surrounding each DNA marker for properties that inhibit or promote the binding of DNA to histones. Our results showed that the majority of markers analyzed were likely to be bound as nucleosomes. Selection of the markers that are more protected to form a multiplex could increase the chance of obtaining a better balanced, easier to interpret DNA profile from degraded samples.

Keywords: Nucleosome, Nucleosome forming potential, Nucleosome positioning signal, Forensic, STR

Back to Article Outline

1. Introduction 

Recent advancements in forensic DNA analysis have focused on improving analysis techniques, such as pyrosequencing [1], increased PCR cycles [2], post-PCR purification [3], and mini-STR designs [4]. These improvements have proved to be successful in obtaining better DNA profiles with degraded DNA samples often found in mass disasters and samples exposed to the environment. However, the intrinsic structural properties of DNA that might prevent its degradation have been overlooked. Using these structural properties as guidelines, forensic scientists might be able to choose the loci that can better withstand degradation and hence obtain more information from a degraded sample.

The binding of the octameric histone cores to 147bp of DNA is a complex, multifactorial process that limits the interactions of DNA with other proteins. The formation and location of nucleosomes, the association of DNA with histones, are known to depend on the following factors: dinucleotide periodicity, base stacking, GC content, and chromatin remodelers [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16]. It has been shown that certain properties, such as low deformation energy [17] and periodicity repeats of GG/CC dinucleotides [9], favor nucleosome formation. They are called “nucleosome positioning signals” [11].

Dixon et al. [18], [19] suggested that a nucleosome could offer protection to the 147bp of DNA that are bound to it from the attacks of endonucleases, which would freely digest post-mortem DNA at exposed sites. An in silico whole human genome annotation for nucleosome exclusion regions also showed that regions free of nucleosomes correlated well with DNase I hypersensitive sites, from which an inference can be made that DNA bound in the nucleosomes could be protected against DNases [20].

The in silico study presented here was carried out to evaluate the “nucleosome forming potentials” (NFPs) (how likely it is for a certain sequence of DNA to be bound by nucleosomes) of 60 forensically important markers (58 STRs plus amelogenin X and Y). After analysis of the softwares available, we explored two nucleosome positioning signals – DNA bendability based on known stiff sequences and dinucleotide base stacking – via two freely available tools, NXSensor [21] and nuScore [22], respectively.

NXSensor searches for three sequences that are known to be rigid and therefore resist bending into a nucleosome. These sequences, when located near each other, could indicate a nucleosome-free region of DNA [21]. A modified version of NXSensor has been shown in silico to achieve good correlations with regions lacking nucleosomes [20]. On the other hand, nuScore works by determining the energy needed to bend a sequence of DNA. This deformation energy is calculated based on the specific arrangements of dinucleotides and their interactions, a phenomenon called dinucleotide stacking. The six possible interactions between the neighbouring two bases are tilt and shift (x-axis), roll and slide (y-axis), and twist and rise (z-axis) [23]. Locations of minimal deformation energy have been shown to correspond well to empirically determined locations of nucleosome dyads, the center of the nucleosome [22].

We hypothesized that some forensically important STR loci evaluated in this study may be more protected by nucleosomes than other loci. Determining which loci are protected could allow them to be incorporated into future forensic identification kits, resulting in a higher discrimination power for certain degraded sample types (saliva, bone, and decomposed remains) than with current profiling methods.

Back to Article Outline

2. Materials and methods 

2.1. Selecting markers and obtaining base sequences 

Fifty-eight STR markers and amelogenin X and Y, totalling 60 markers (Table 1), were selected based on their past use and current recommendations by the forensic community. Sequences were obtained from the NCBI Human Genome Map (http://www.ncbi.nlm.nih.gov/genome/guide/human/). These sequences were center-aligned at the tandem repeat units and truncated for 200bp at both the 5′ and 3′ end, yielding a sequence of 400 bases.

Table 1. The 58 STR markers plus amelogenin X and Y, their GenBank accession number and chromosomal position in the latest human GRCh37 assembly.

2.2. NXSensor mechanisms and parameters 

The algorithm of Nucleosome eXclusion Sensor (NXSensor version 1.3.1) (http://www.sfu.ca/∼ibajic/NXSensor/) reads an input sequence for three known nucleosome-free sequences: 10 bases of poly-A, 10 bases of poly-T, and a combination of Gs and Cs (A_≥10, T_≥10, or [(G/C)₃N₂]_≥3). If any of these sequences are found, the program outputs the sequence in FASTA format and highlights the nucleosome-free region.

All 60 markers were evaluated and accessibility scores were given as a measure of how accessible the input sequence was to DNA-binding proteins. The score was calculated using the following formula:

where A the accessibility score; L_o the total length of open contiguous segment; OS_min the minimum length of open segment; L_i the length of input sequence; and L_a the total length of ambiguous segments. An accessibility score of 0 indicated the whole input sequence contains no sequence that inhibits nucleosome formation while a score of 1 indicated the whole sequence is open for access by proteins and is not bound as a nucleosome.

The default settings used were: 147bp window size; minimum number of exclusion sequences considered significant=1; and the minimum length of open segments=10.

A marker was deemed to have a high NFP if the accessibility score was close to zero and a low NFP close to one.

2.3. NuScore mechanisms and parameters 

NuScore (http://compbio.med.harvard.edu/nuScore/) was used to evaluate the DNA deformation energy based on dinucleotide stacking properties—tilt, shift, roll, slide, twist, and rise. Randomized sequences were generated 100 times with the same dinucleotide content as the input sequence. The program options selected were: template 2cv5 (human); best of two orientations; and 164bp window size. Two output values – DNA deformation energy and nucleosome positioning score (NPScore) – were used in this study.

The DNA deformation energy measures the amount of energy required to impose the structure of the nucleosome bend onto the input sequence; whilst the NPScore shows the significance in deviation of the deformation energy at one point from its neighbouring positions. Supplementary materials from [22] used an NPScore threshold of less than or equal to −2 to indicate a possible nucleosome dyad location, and the same threshold was applied in this study. A more stringent threshold of −3 was evaluated in this study as well.

2.4. Comparisons of original base sequences with random arrangements 

All 60 markers were compared with random sequences of the same dinucleotide content to determine if the positioning of the nucleosome dyad is dependent upon the specific arrangements of dinucleotides and their interactions. The number of locations with an NPScore more negative than two thresholds (−2 and −3) were counted and compared statistically.

A marker was deemed to have a high NFP when the locations of NPScore crossing the threshold (NPScore≤−2) were high and vice versa.

Back to Article Outline

3. Results 

3.1. NXSensor 

An overall median of 0 (fully accessible) and a standard deviation of 0.019 indicated that the majority (81.7%) of markers tested had high NFP and hence were more probable to associate with histones to form nucleosomes. Ten out of the 60 markers (18.3%) contained short nucleotide sequences that were deemed “stiff” and were less probable to exist as nucleosomes. Their accessibility scores ranged from 0.028 to 0.098 (Fig. 1).

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1. 

  Accessibility scores (calculated using Eq. (1)) of 11 STR loci with nucleosome exclusion sequences.

3.2. NuScore 

An example graphical output of an NPScore profile of D18S51 is shown in Fig. 2a. The alternating high-low score seen in the figure was typical of every sequence. The minima signified locations where there was potential for a nucleosome dyad to exist. In this profile, a reference line is shown at −2, as suggested by [22]. Values below −2 and −3 were counted for each STR locus. All loci displayed at least one possible location for a nucleosome dyad (threshold of −2) in the 400 bases input (Table 2). The medians of possible nucleosome dyad locations were 7 and 1 for the threshold of −2 and −3 respectively, with standard deviations at 2.380 and 0.851. The markers with the highest potential dyad locations of 12 were D21S11 and D10S1435, meaning that these two loci were the most likely to be bound to nucleosomes.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 2. 

  Nucleosome positioning score profile of D18S51 (a.u.=arbitrary units) with reference line at −2: (a) original base sequences and (b) random dinucleotide profile of the same composition.

Table 2. Number of possible locations for a nucleosome dyad with the threshold of −2 and −3 for the 60 markers tested. Original (Ori) and random (Ran) indicate the arrangement of the base sequences. Original arrangement is found in a human genome and random is the same dinucleotides arbitrarily rearranged. The table is arranged in descending order of possible locations for Ori in threshold −2.

Within the central 100bp, the markers were divided into three groups based on their scores. Group A comprised 27 markers with scores from 0 to 2, i.e. there were two or less positions in the central 100bp that crossed the threshold of −2. Group B comprised 28 markers with scores between 3 and 5, inclusive, and group C comprised 5 markers whose scores were above 6 (Table 3).

Table 3. 60 markers divided into three groups according to the number of positions crossing the threshold of −2 at the central 100 bases (sorted in alphabetical order).

3.3. Comparisons of original base sequences with random arrangements 

A set of 100 random sequences with the same dinucleotide composition was generated by the nuScore program for each marker. The number of possible dyad locations for both arrangements (original and random) and both thresholds (−2 and −3) are listed in Table 2. For example, using D18S51 as an input sequence, a random sequence profile was generated and displayed for direct comparison with the original profile (Fig. 2b).

Statistical comparison of original and random configurations containing dyad locations with a threshold of −2 gave a p-value of 0.004, indicating a significant difference in the scores obtained from the two different configurations. When the threshold was set to −3, no significant difference was observed between the two configurations (p=0.466). Due to the low number of positions crossing the threshold at −3 (Table 2), threshold −2 was chosen for further experiments.

Comparison of accessibility scores from NXSensor to threshold −2 and −3 scores for each STR locus yielded correlation coefficients of 0.024 and −0.13, respectively. This observation revealed that there was no linear correlation of accessibility scores to NPScores.

Back to Article Outline

4. Discussion 

Two hundred base pairs to both the 5′- and 3′-end from the center of the repeat units were used for the sequence analysis described in this study. This total of four hundred base pairs was chosen because the four hundred base pairs unit wholly encompasses the repeat motifs of the markers as well as the possible primer binding sites flanking the motifs. Moreover, the largest loci of the widely used commercials kits do not generally go beyond this size, albeit with a few exceptions, such as Penta E of PowerPlex^® 16 (size range 379–474bp) [24]. In addition, most STRs in current use do not have repeat units (without the flanking regions) that exceed 147bp, which is equivalent to 36–37 repeat units for a tetranucleotide STR. The reason that amplicon lengths of commercial kits such as SGM Plus™ extend much greater than the repeat unit size is because of the flanking regions for convenient primer design and multiplexing. Hence, using mini-STR primers as shown in [25] as an example, the actual amplicon length of the markers can be reduced to less than the nucleosomal protection size of 147bp.

Given that STRs have varying number of repeats depending on each individual and that our methods center-aligned the sequence, the flanking regions will change accordingly with each allele. This could have an effect on the NFPs. However, center-aligning the sequences was deemed important because, in theory, the closer the nucleosome dyad is to the center of the repeat units, the higher the chance that the primer binding sites and the repeat units would remain intact after DNA degradation (due to nucleosome protection) and that successful PCR amplification could occur.

Based on the NXSensor results, the markers were divided into two groups—ones with exclusion sequences and ones without. The validity of nucleosome protection conferred upon the markers could be empirically determined by designing primers for these ten loci, particularly D9S1122, with the highest accessibility scores (Fig. 1), and then compare their performance on artificially degraded DNA and casework samples with a marker that had no exclusion sequence, e.g. vWA. If the hypothesis proposed in this study is correct, the ten loci with accessibility scores of more than 0 should exhibit properties associated with degraded DNA and/or low-template DNA amplification [2] while these effects should be dampened with the other 49 loci.

The nuScore results showing at least one potential nucleosome dyad location (NPScore≤−2) within the 400bp of each marker were expected. Nucleosomes serve to facilitate compacting of the chromatin for higher order structure [14] and, as a nucleosome binds approximately 147bp of DNA, at least one dyad should be seen in a sequence as long as 400bp. The core 100bp was more important as discussed above and was used to categorize the STR loci. The markers in group A and group C could be targeted for further empirical comparisons for evidence that may validate the protective capabilities of nucleosomes on STR loci.

The numbers of possible dyad location of the 60 markers in their original arrangements were compared to those of the random arrangements in order to determine if these locations are influenced by the specific arrangement of dinucleotides and their interactions. The change in the locations of the minima (Fig. 2) and the statistical differences between the two suggest that both the numbers of possible dyad locations and the positions they might take up depend on the arrangement of the bases and not the nucleotide content.

The lack of correlation between the two programs may be due to their basic designs. The NXSensor program searches for sequences that are inhibitory to nucleosome binding based on known strong inhibitory signals [20] while the nuScore program evaluates the dinucleotide stacking properties of the input sequence. Nucleosome binding and attraction is a multifactorial event, with commonly known variables being dinucleotide periodicity and stacking, GC content and chromatin remodelers [9], [11], [12], [13], [15], [20], [26]. Since this attraction depends on more than one single factor, a significant difference might not be observed when only one or two factors are considered as in our study. Furthermore, these signals only indicate the probability of finding a nucleosome at a given location and not experimentally mapped nucleosome positions. Therefore further experiments using artificial degradation of saliva samples are being carried out to validate our in silico findings.

Also, all the STR loci evaluated in these studies, with the exception of amelogenin, were located within an intron or intergenic region. Our findings agree with Vinogradov [27], who showed that these regions are enriched with NFPs because they do not have to be regulated, transcribed (for intergenic STRs), or translated (for intron STRs). Therefore they are less likely to be accessed by proteins.

Other nucleosome prediction programs are also available online for research use. The most recent program released in December 2009 is “FineStr” (http://www.cs.bgu.ac.il/∼nucleom/) [28], which is based on Caenorhabditis elegans universal nucleosome positioning pattern [29], [30]. Another noteworthy program based on discriminant analysis of dinucleotide frequency is called “Recon” (http://wwwmgs.bionet.nsc.ru/mgs/programs/recon/) [31] and it has been available since 2001. Given the relatively fast nature of the field, we decided to use NXSensor and nuScore, which are the two programs that are freely available and based on most up-to-date data when the study was carried out.

Back to Article Outline

5. Conclusion 

Assessment of NFPs of forensically important STRs could be beneficial as nucleosome-bound DNA is less likely to degrade due to being in the ‘bound configuration’. Our findings show that most STR loci in use nowadays are already somewhat protected from degradation by nucleosomes. Selecting the loci with stronger attraction to nucleosomes for a multiplex could increase the chance of obtaining a better balanced profile with fewer allelic drop-outs. Further work using time-series degraded saliva samples will be employed to empirically determine the differences between bound and unbound STRs under in vitro conditions.

Back to Article Outline

References 

1. Anna-Maria D, Hanna E, Marie A. Forensic analysis of autosomal STR markers using Pyrosequencing. Forensic Sci. Int. Genet. 2010;4:122–129
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (688 KB)
   • CrossRef
2. Gill P, Whitaker J, Flaxman C, Brown N, Buckleton J. An investigation of the rigor of interpretation rules for STRs derived from less than 100pg of DNA. Forensic Sci. Int. 2000;112:17–40
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (1093 KB)
   • CrossRef
3. Smith PJ, Ballantyne J. Simplified low-copy-number DNA analysis by post-PCR purification. J. Forensic Sci. 2007;52:820–829
   • View In Article
   • MEDLINE
   • CrossRef
4. Butler JM, Shen Y, Mccord BR. The development of reduced size STR amplicons as tools for analysis of degraded DNA. J. Forensic Sci. 2003;48:1054–1064
   • View In Article
   • MEDLINE
5. Ercan S, Lieb JD. New evidence that DNA encodes its packaging. Nat. Genet. 2006;38:1104–1105
   • View In Article
   • MEDLINE
   • CrossRef
6. Ioshikhes I, Bolshoy A, Derenshteyn K, Borodovsky M, Trifonov EN. Nucleosome DNA sequence pattern revealed by multiple alignment of experimentally mapped sequences. J. Mol. Biol. 1996;262:129–139
   • View In Article
   • MEDLINE
   • CrossRef
7. Ioshikhes IP, Albert I, Zanton SJ, Pugh BF. Nucleosome positions predicted through comparative genomics. Nat. Genet. 2006;38:1210–1215
   • View In Article
   • MEDLINE
   • CrossRef
8. Kiyama R, Trifonov EN. What positions nucleosomes?—a model. FEBS Lett. 2002;523:7–11
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (294 KB)
   • CrossRef
9. Kogan SB, Kato M, Kiyama R, Trifonov EN. Sequence structure of human nucleosome DNA. J. Biomol. Struct. Dyn. 2006;24:43–48
   • View In Article
   • MEDLINE
10. Lowary PT, Widom J. New DNA sequence rules for high affinity binding to histone octamer and sequence-directed nucleosome positioning. J. Mol. Biol. 1998;276:19–42
    • View In Article
    • MEDLINE
    • CrossRef
11. Peckham HE, Thurman RE, Fu Y, Stamatoyannopoulos JA, Noble WS, Struhl K, et al. Nucleosome positioning signals in genomic DNA. Genome Res. 2007;17:1170–1177
    • View In Article
    • CrossRef
12. Rando OJ, Ahmad K. Rules and regulation in the primary structure of chromatin. Curr. Opin. Cell Biol. 2007;19:250–256
    • View In Article
    • MEDLINE
    • CrossRef
13. Segal E, Fondufe-Mittendorf Y, Chen L, Thastrom A, Field Y, Moore IK, et al. A genomic code for nucleosome positioning. Nature. 2006;442:772–778
    • View In Article
    • CrossRef
14. Kornberg RD, Lorch Y. Twenty-five years of the nucleosome, fundamental particle of the eukaryote chromosome. Cell. 1999;98:285–294
    • View In Article
    • MEDLINE
    • CrossRef
15. Yuan GC, Liu YJ, Dion MF, Slack MD, Wu LF, Altschuler SJ, et al. Genome-scale identification of nucleosome positions in S. cerevisiae. Science. 2005;309:626–630
    • View In Article
    • CrossRef
16. Bettecken T, Trifonov EN. Repertoires of the nucleosome-positioning dinucleotides. PLoS ONE. 2009;4:e7654
    • View In Article
    • CrossRef
17. Travers AA. The structural basis of DNA flexibility. Philos. Trans. Royal Soc. London. Series A: Math. Phys. Eng. Sci. 2004;362:1423–1438
    • View In Article
18. Dixon LA, Murray CM, Archer EJ, Dobbins AE, Koumi P, Gill P. Validation of a 21-locus autosomal SNP multiplex for forensic identification purposes. Forensic Sci. Int. 2005;154:62–77
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (1255 KB)
    • CrossRef
19. Dixon LA, Dobbins AE, Pulker HK, Butler JM, Vallone PM, Coble MD, et al. Analysis of artificially degraded DNA using STRs and SNPs—results of a collaborative European (EDNAP) exercise. Forensic Sci. Int. 2006;164:33–44
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (847 KB)
    • CrossRef
20. Radwan A, Younis A, Luykx P, Khuri S. Prediction and analysis of nucleosome exclusion regions in the human genome. BMC Genom. 2008;9:186–1186
    • View In Article
21. Luykx P, Bajic IV, Khuri S. NXSensor web tool for evaluating DNA for nucleosome exclusion sequences and accessibility to binding factors. Nucleic Acids Res. 2006;34:W560–565
    • View In Article
    • CrossRef
22. Tolstorukov MY, Choudhary V, Olson WK, Zhurkin VB, Park PJ. nuScore: a web-interface for nucleosome positioning predictions. Bioinformatics. 2008;24:1456–1458
    • View In Article
    • CrossRef
23. Dickerson RE. Definitions and nomenclature of nucleic acid structure parameters. J. Biomol. Struct. Dyn. 1989;6:627–634
    • View In Article
    • MEDLINE
24. Technical Manual: PowerPlex® 16 System #TMD012, Promega Corporation, 2008.
    • View In Article
25. Hill CR, Kline MC, Coble MD, Butler JM. Characterization of 26 miniSTR loci for improved analysis of degraded DNA samples. J. Forensic Sci. 2008;53:73–80
    • View In Article
    • CrossRef
26. Trifonov EN. Curved DNA. CRC Crit. Rev. Biochem. 1985;19:89–106
    • View In Article
    • MEDLINE
27. Vinogradov AE. Noncoding DNA, isochores and gene expression: nucleosome formation potential. Nucleic Acids Res. 2005;33:559–563
    • View In Article
    • CrossRef
28. Gabdank I, Barash D, Trifonov EN. FineStr: a web server for single-base-resolution nucleosome positioning. Bioinformatics. 2010;btq030
    • View In Article
29. Trifonov EN. Nucleosome positioning by sequence, state of the art and apparent finale. J. Biomol. Struct. Dyn. 2010;27:741–746
    • View In Article
30. Trifonov EN. Base pair stacking in nucleosome DNA and bendability sequence pattern. J. Theor. Biol. 2010;263:337–339
    • View In Article
    • CrossRef
31. Levitsky VG. RECON: a program for prediction of nucleosome formation potential. Nucleic Acids Res. 2004;32:W346–349
    • View In Article
    • CrossRef