Figures
Abstract
High-throughput sequencing continues to produce an immense volume of information that is processed and assembled into mature sequence data. Data analysis tools are urgently needed that leverage the embedded DNA sequence polymorphisms and consequent changes to restriction sites or sequence motifs in a high-throughput manner to enable biological experimentation. CisSERS was developed as a standalone open source tool to analyze sequence datasets and provide biologists with individual or comparative genome organization information in terms of presence and frequency of patterns or motifs such as restriction enzymes. Predicted agarose gel visualization of the custom analyses results was also integrated to enhance the usefulness of the software. CisSERS offers several novel functionalities, such as handling of large and multiple datasets in parallel, multiple restriction enzyme site detection and custom motif detection features, which are seamlessly integrated with real time agarose gel visualization. Using a simple fasta-formatted file as input, CisSERS utilizes the REBASE enzyme database. Results from CisSERS enable the user to make decisions for designing genotyping by sequencing experiments, reduced representation sequencing, 3’UTR sequencing, and cleaved amplified polymorphic sequence (CAPS) molecular markers for large sample sets. CisSERS is a java based graphical user interface built around a perl backbone. Several of the applications of CisSERS including CAPS molecular marker development were successfully validated using wet-lab experimentation. Here, we present the tool CisSERS and results from in-silico and corresponding wet-lab analyses demonstrating that CisSERS is a technology platform solution that facilitates efficient data utilization in genomics and genetics studies.
Citation: Sharpe RM, Koepke T, Harper A, Grimes J, Galli M, Satoh-Cruz M, et al. (2016) CisSERS: Customizable In Silico Sequence Evaluation for Restriction Sites. PLoS ONE 11(4): e0152404. https://doi.org/10.1371/journal.pone.0152404
Editor: Manoj Prasad, National Institute of Plant Genome Research, INDIA
Received: July 20, 2015; Accepted: March 14, 2016; Published: April 12, 2016
Copyright: © 2016 Sharpe et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The Nostoc sp. PCC 7107 genome is available from NCBI Accession number NC_019676 www.ncbi.nlm.nih.gov/nuccore/NC_019676.1. The Arabidopsis EST dataset is available from TAIR ftp://ftp.arabidopsis.org/home/tair/Sequences/ATH_cDNA_EST_sequences_FASTA/. Other relevant data are within the paper and its Supporting Information files.
Funding: This work was supported by WSU Agriculture Research Center Hatch funds to AD and KE. US Department of Agriculture National Research Initiative (USDA-NRI) grant 2008 -35300-04676 to AD and AK supported AH and JG. TAK acknowledges the National Institutes of Health Protein Biotechnology Training Grant T32GM008336 and ARCS Fellowship Program. RMS's work was supported by the National Science Foundation under Grants IOS 0641232 and MCB 1146928 and by Civilian Research and Development Foundation Grant RUB1-2982-ST-10 to Gerald E. Edwards, WSU. Work related to characterization of the cfq mutant was supported by US Department of Energy, Office of Science, Basic Energy Sciences Program DE-FG02-04ERl5559 to DMK.
Competing interests: The authors have declared that no competing interests exist.
Introduction
High-throughput sequencing technologies continue to generate vast amounts of information. The DNA sequence information is processed for quality and assembled into contigs resulting in the generation of mature sequence data that is subsequently utilized by biologists in wet lab experiments. Availability of user-friendly computational tools specifically created to process large quantities of sequence information from multiple samples that catalyze the translation of these countless data into useful knowledge for addressing biological questions remains a bottleneck. Existing sequence data can be harnessed for nucleotide polymorphism information, ascertaining genetic diversity in a population, and reduced representation sequencing. The consequences of nucleotide polymorphisms are diverse. They might result in altering the phenotype if there is a change in an amino acid or alterations in the regulatory regions. Alternatively, these may be inconsequential mutations. Biologists endeavor to first identify and then utilize the polymorphic information to establish causal relationships between the genotype and the phenotype in genomics and genetics approaches.
There are several approaches in use that exploit nucleotide polymorphism information. On a global genomic scale, nucleotide polymorphism information is generated using whole genome sequencing, reduced representation sequencing [1–3], genotyping by sequencing [4,5], and SNP arrays [6,7]. Genotyping by sequencing and reduced representation sequencing both utilize restriction site information during sequencing library preparation for genomes [1, 2] and transcriptomes [3–5]. An example is Restriction-site Associated DNA (RAD) sequencing that enables identification of polymorphisms which are subsequently used as DNA markers for population analysis [6, 7]. Whole genome analysis of restriction sites can provide better information to help guide decisions for enzyme selection in digesting DNA for RAD sequencing libraries as well as BAC library production and sub-cloning. Transcriptome analysis via sequencing of 3’ untranslated regions (3’UTR) of cDNA libraries was enabled through the utilization of restriction enzyme digests in maize [3] and sweet cherry [4].
The afore-mentioned approaches are overkill when working with a single or a few hundred genes as is typical of most projects. For such focused applications, methods such as high resolution melting [4], allele specific PCR [8], single locus restriction fragment length polymorphisms (RFLP), amplified fragment length polymorphisms (AFLP) and cleaved amplified polymorphic sequences (CAPS) [9] are used. Of these, methods based on restriction enzyme digestion are relatively easy, widely used, reproducible, and cost-effective to perform and analyze due to the reduced need for specialized equipment and expertise. For genetics applications, restriction enzyme based molecular markers are commonly employed. Historically, these molecular markers have been developed either through trial and error or from polymorphisms among a limited set of individuals. With the availability of high-throughput sequencing technologies, the onus has shifted to identifying multiple, site-specific polymorphisms across large populations. CAPS markers, also known as PCR based Restriction Fragment Length Polymorphic markers (PCR-RFLP), are routinely used for mapping traits in populations and for enabling efficient breeding [9]. CAPS markers are popular due to their relatively low cost and general ease of use through the reliance on the common, simple molecular biology tools of PCR, enzymatic digests and gel electrophoresis [10]. These strategies, however, rely on a priori knowledge including the location and sequence of the restriction sites. In addition to polymorphism screening, many types of molecular biology methods utilize restriction site information and therefore require an efficient tool to analyze large sequence datasets.
Several current restriction site analysis tools, summarized in Table 1, have been designed to handle one or several small sequences for targeted analysis [11–17], not the thousands or millions that are now available with high-throughput sequencing. NEBcutter is the most generalized restriction site analysis tool with the others focusing primarily on developing DNA markers. NEBcutter provides useful functionalities from cloning analysis to gel predictions based on different types of gels [17]. Version 2.0 of this web tool, at its farthest limit, handles a single sequence of less than 300 kbases or an input file of a single sequence less than 1 Mbyte for analysis regarding any number of enzymes from the REBASE database [8]. The analysis pipeline enables most molecular functionalities but is not set up for high-throughput analysis nor multiple sequence analysis.
A comparison of some essential traits of CisSERS and other restriction site analysis programs highlights the advantages of CisSERS and some of the shared components with previously available tools. Many of these tools were designed for CAPS marker or derived CAPS (dCAPS) marker development and each has varying limitations.
Like NEBcutter, most of the previous molecular marker developing tools are web-server based, limiting functionality for high-throughput analysis depending on the users’ internet connection and the tool’s server availability. The need to upload large datasets to these webserver-based programs can cause a significant bottleneck. Since the molecular marker development tools are mainly used to design CAPS markers, several of these tools have added primer design for amplification of a region around a polymorphism-modified restriction site. Additionally, several of these tools include levels of automated decision making that aids primer set and enzyme selection although this reduces user control and preferences.
Here we present a novel tool, CisSERS: Customizable in silico Sequence Evaluation for Restriction Sites that was developed to enable high-throughput analysis of mature multiple sequences for restriction sites with an embedded dynamic visualization functionality when identifying and selecting restriction enzymes or custom motifs for subsequent wet-lab applications. CisSERS output includes DNA digest information including an agarose gel prediction, to facilitate the user’s decision making process for selection of the most appropriate enzyme(s) for the project application. Unlike any other program in its class, CisSERS allows for custom motif detection to identify conserved sites, such as cis-acting elements or trans-acting protein binding sites, among all sequences and it even predicts amplicon lengths when oligonucleotides are input as custom motifs. For user convenience and project efficiency, CisSERS retains project files to provide easy access to the predicted cut site information to reduce time when the project requires iterative interactions, additional analysis or when two projects require comparison. In summary, CisSERS is expected to bridge the gap between sequence acquisition and implementation of diverse wet-lab approaches in order to address biological questions.
Material and Methods
Overview of CisSERS
CisSERS was developed as a standalone program to provide processing of fasta files for restriction site and custom motif analyses generating tables of counts and predicted gel image as outputs. CisSERS is a java based graphical user interface built around a perl backbone presented in a standalone java execution file. CisSERS requires a onetime download of latest release of Java Runtime Environment (JRE) and Perl and, the CisSERS java archive (jar). While there are no known platform dependencies, the program runs without any problems in JRE version 7 on all operating systems. Using hundreds of thousands or millions of sequences or a single sequence in a single fasta file, selected motifs are identified, displayed and analyzed through a java based graphic user interface (GUI). The fundamental string matching functionality embedded in PERL enables motif identification. After analysis is complete, the outputs are displayed in the program including tables describing the cut counts and locations for each restriction site and dynamically created predicted gel images. Fig 1 provides a graphical overview of the CisSERS workflow. The program (CisSERS.jar), User manual for the CisSERS program and CisSERS Overview and Usage graphics can be found in S1, S2 and S3 Files. The CisSERS source code is available in S5 File.
A graphical depiction of the three phases of the CisSERS process and their subsections.
General Workflow
User Input
For data entry into CisSERS, the user selects the fasta-formatted DNA sequence file for processing using a standard folder/file selection. The file is analyzed to verify that it is in the proper format and CisSERS will warn the user if irregularities are identified once it is started. For transcriptome analysis approaches where poly-A trimming and 5’ to 3’ orientation is desired, selection of the ‘Sequences have Poly-A tails’ check box enables this pre-processing (refer to the User’s Manual for all program option locations). If poly-A is selected, identification and subsequent trimming of the poly-A is completed. Sequences without a recognized poly-A sequence are placed in a separate file and are not processed further. The user can also determine how far into a sequence the poly-A tail search will continue. Poly-A tail identification, based on the approach found in the EMBOSS program ‘trimest’ [19], occurs if 4 or more consecutive A’s are found within the specified range. The poly-A tail search will be extended until more than 1 non-A character is identified. If a search limit has been specified, the algorithm will extend the poly-A tail beyond the range to most effectively trim the sequence. While the poly-A identification is running, the algorithm is also applied to identify poly-T heads which are then reverse complemented and placed in the proper 5’ to 3’orientation for all further analyses. The trimmed files, containing all sequences in the 5’ to 3’ orientation, and non-trimmed files generated during this process are saved for further use if desired. Without the Poly-A selection, all reads are processed in their input orientation and must be oriented in the 5’ to 3’ direction for use in non-transcriptomic approaches.
The second step of user input is to choose the desired ‘cut site’ area. The user can choose the area of predicted cut sites from either end or along the entire length of the sequence. The cut sites are predicted using pattern or expression matching functionality inherent in Perl. Fasta files provided by the user are opened and each sequence line is read as a variable to which the master list of enzymes available from the REbase database [17] are identified using the regular expression matching functionality. The 3’ and 5’ options enable the selection of a range of the sequence to be analyzed from the desired end. The user enters this information by dragging the slider or entering the exact positional information in the dialog boxes below the slider diagram.
The third and final input required prior to initiating analysis by CisSERS is the enzyme selection. The master list of enzymes is retrieved from the REbase database [17] and can be periodically updated by the user from within CisSERS if the host computer has internet access. Enzyme selection can be done through a checkbox tree, a filtering window, or through a name/site search. This allows the user to choose enzymes meeting a number of criteria including custom lists. User defined enzymes and recognition sequences can also be added to the database if desired. The options menu also allows the user to select which outputs are desired prior to starting the analysis.
Analysis
After all user inputs are entered, selecting ‘Run’ at the bottom of the screen begins the analysis process. During this time, a ‘processing’ tab is shown depicting the progress of the analysis through each stage. Once the analysis is complete, this tab will disappear and ‘Summary’, ‘Best’ and ‘Top’ tabs will appear when the ‘Sequences have Poly-A tails’ box is selected; when the ‘Sequences have Poly-A tails’ box is not selected, an additional ‘Gel Visualization’ tab appears.
Results/Outputs
The primary outputs of CisSERS are: Summary, Best, Top tables and Predicted Gel Visualization as mentioned previously. Each of these outputs is shown on an individual tab and can be saved individually by CisSERS. Projects can be saved and reloaded to eliminate the need to reprocess the data.
The ‘Summary’ table displays the total results for each enzyme. This includes the total number of sequences which contain at least one occurrence of that enzyme’s recognition sequence, the total number of cut sites in the fasta file and the percentage of sequences that are cut by this enzyme. The ‘Best’ table is used for finding combinations of enzymes that cut the most sequences. For 3’UTR sequencing, a minimum number of enzymes that cut nearly every transcript in the desired range are ideal. Using a greedy approach, where sequences cut by the best enzyme are removed and the next best enzyme is identified, the enzymes are listed with the combined percentage of sequences cut. Additionally, since there are minimum lengths for processing DNA through some applications, the 'Best' table also displays cut sites in the 'Pre-cut Area', the area between the beginning of the sequence and the beginning of the desired cut area. The last table produced is the ‘Top’ table. This table is a filtered version of the ‘Summary’ table that only displays the enzymes cutting a minimum percentage of the sequences. This setting defaults to 95% and is adjustable through the options menu. These data tables combine to inform the user of the restriction site information necessary to enable many biological approaches.
Cut Site Identification and Gel Visualization
Basic functionalities of CisSERS were tested by evaluating its ability to properly identify restriction sites or motifs using expression matching functionality in Perl and produce a predicted gel image on datasets constructed with known restriction enzyme recognition site sequence inserted into the middle of poly-T sequences resulting in unique sequences with a final length of 60 bases. The set of formulas CisSERS uses to produce fragment size to distance relationships were derived and modified from previous reports [9–12]. Briefly, the published formulae describe fragment to fragment interaction in a resistive manner and thus different sized DNA fragments have a size specific differential resistance associated with them while passing through a matrix of agarose fragments. These size specific differential resistance associations were applied to Ohm’s Law, which states that current multiplied by resistance equals voltage (IR = V), and using a voltage constant of 70 VDC to obtain size specific current values. The size specific current values were extrapolated as velocity values and, utilizing a time constant of 45 minutes, distance relationships were obtained using the velocity multiplied by time equals distance equation (VT = D).
Custom Motif Identification
The ability to examine a certain set of sequences for shared motifs has vast applications. These motifs could be cis-acting transcriptional or translational regulators that impact gene expression and ultimately the phenotype of the organism. CisSERS offers this distinctive feature to search for custom motifs enabling such analyses. Custom motifs are input by the user and added to the Cut Sequence list. They are selected via the check box option and are identified in the sequence of interest based on the regular expression matching functionality in Perl.
Demonstration case 1.
The motif detection feature of CisSERS was validated by analyzing the Nostoc sp. PCC 7107 genome. Nostoc is a commonly used nitrogen fixing microbe used in undergraduate curriculum and possesses a circular genome. The Nostoc genes may start with a Pribnow box or an alternate AT rich version that corresponds to the IUPAC code WWWWWW motif where W corresponds to either an A or a T. This motif is generally located between the -4 and -12 position relative to the ATG start codon. The number of bases between the motif and the start codon was varied by inserting 4–12 N’s, WWWWWWN4-12 yielding the motifs shown in Table 2. Annotated genes with these motifs were identified and analyzed for the subset of genes that contains all 9 motifs in the transcriptional initiation area.
Demonstration case 2.
Typically polyadenylation of mRNA 3’ untranslated region (UTR) requires a polyadenylation initiation site that facilitates the binding of the cleavage/polyadenylation specificity factor (CPSF) complex which cleaves and polyadenylates the 3’ end. The majority of eukaryotic mRNA transcripts possess a polyadenine (polyA) tract on the 3’ end of the transcript. The polyA tract has been implicated in regulation of mRNA degradation and translation [13]. Alternative processing of mRNA transcripts can lead to different isoforms of a gene either performing alternative functions, differential regulation in a pathway or gene auto-regulation or non-functionality of a gene product [14]. Among the different forms of alternative processing is the existence of premature polyadenylation of a transcript. CisSERS was used to find the predicted polyA initiation sites within 300 bases of the terminal 3’ reported base for each of the ESTs contained in the TAIR ATH_cDNA_EST_sequences_FASTA file for transcripts that could support this form of alternative mRNA processing. Due to the large memory requirements to process the complete file, the ATH_cDNA_EST_sequences_FASTA file was subdivided into 37 datasets, 36 files containing 50,000 fasta sequences and 1 containing 16,638 sequences, for processing by CisSERS and the individual subset results were collated. The human canonical AATAAA polyA initiation recognition site, as well as previously identified eukaryotic polyA initiation recognition sites [15], were used as input motifs to identify polyA initiation recognition sites present in the expressed sequence tags of the Arabidopsis dataset (Table 3).
CAPS marker development
CAPS marker development is an important feature of CisSERS that was tested to verify the utility of this function. CAPS markers were developed for diploid and polyploid species by analyzing the input sequences in demonstration cases 1 and 2 with CisSERS using all restriction enzymes. The restriction digestion results were visualized using the virtual gel output and digestions patterns were visually parsed for discernible differences in sizes of the digested DNA fragments with each restriction enzyme. A single restriction enzyme or a combination of preferably two enzymes can be used to obtain different restriction digestion pattern from similar sequences with embedded polymorphisms, thus resulting in the development of a CAPS marker. As NEBcutter V2.0 does not have a method for analyzing multiple sequences in parallel, which is critical for enzyme comparison, only CisSERS-identified enzymes for CAPS markers were analyzed in subsequent biological experiments.
Demonstration case 1.
ATPC1 is one of the two nuclear encoded genes in Arabidopsis for the γ subunit of the chloroplast ATP synthase [17]. The coupling factor quick recovery (cfq) mutant of Arabidopsis was identified as a point mutation in the ATPC1 gene and reduces overall photosynthetic capabilities [18]. The sequences of wild type Arabidopsis ATPC1 and the cfq mutant form were processed through CisSERS. The purpose for CisSERS analysis was to identify at least one restriction enzyme displaying significant visual differences for use as a CAPS marker to enable population screening. DNA was extracted from wild type, cfq mutant, and heterozygous Arabidopsis plants and the ATPC1 gene was amplified. The product was then digested with TaqI identified by CisSERS for 1 hour at 65°C and electrophoresed on a 10% TBE-Acrylamide gel (Bio-Rad), stained with ethidium bromide, and visualized (Fig 2).
The two were linked to create the “F1 het” lane image while the F1 heterozygous plant DNA was analyzed and labeled “F1 het” in the wet-lab validation image. The banding patterns of all three samples of the CisSERS prediction match the wet-lab validation confirming the effectiveness of CisSERS to determine effective CAPS marker enzymes.
Demonstration case 2.
A gene putatively involved in bitter-pit disorder of apple was identified in previous work (Schaeffer and Dhingra, unpublished). Apple is an allotetraploid with a recently published genome [20]. Cloning and sequencing of this gene from eight apple cultivars varying in degree of disorder prevalence was completed. These sequences were then analyzed with CisSERS to identify an enzyme which separates the major alleles present in these cultivars. Wet-lab evaluation was performed by amplifying the region and digesting with Cac8I for 3 hours at 37°C. The resulting DNA fragments were electrophoresed on a 2% agarose gel and visualized (Fig 3).
A. CisSERS predicted gel image of 12 identified alleles from 8 apple cultivar’s cDNA clones, and 2 linked gel images (Gold_Del, and Red_Grav). B. Wet-lab electrophoresed gel image of amplified products (#a) and corresponding restriction digest (#b); 1. ‘Macintosh’, 2. ‘Winesap’, 3. ‘Red Gravenstein’, 4. ‘Haralson’, 5. ‘Cox’s Orange Pippin’, 6. ‘Braeburn’, 7. ‘HoneyCrisp’, 8. ‘Golden Delicious’, MM = 100bp DNA molecular marker. Analysis of the individual cultivars (A: Haralson 2 and B: 4b) suggest that ‘Haralson’ is homozygous for the sequenced allele; (A: Macintosh 9, Macintosh 2, Macintosh 5 and B: 1b) indicates that each allele present in ‘Macintosh’ is not yet sequenced; and (A: Cox_Org 10, Cox_Org 5 and B: 5b) also indicates that each allele of ‘Cox’s Orange Pippin’ has not yet been sequenced.
High-Throughput Analysis
CisSERS has the unique capability of analyzing large datasets rapidly, which is a major upgrade compared to other restriction site analysis programs. To demonstrate this functionality, the Arabidopsis thaliana EST cDNA dataset was downloaded from the TAIR website and processed with CisSERS. 1,816,638 sequences with an average length of 321 bases, longest sequence of 2,883 bases and the shortest EST of 1 base. To limit the output for Table 4 and to emphasize the customization of enzyme selection, the 6 base cutter restriction enzyme set was used to process the entire dataset.
Results
Custom Motif Identification
Demonstration case 1.
While there are few canonical transcriptional start sites (TATAAT sites) associated with Nostoc genes (Nos7107_0081 hypothetical protein and Nos7101_1087 group 1 glycosyl transferase on the forward strand and Nos7107_3714 hypothetical protein on the reverse strand), the distinctive functionality of CisSERS can be used with the degenerate nucleotide base codes to increase the identification of possible cis-element Pribnow box motif variations. A total of 41,941 potential AT rich transcriptional start sites are present in the Nostoc genome on the forward strand when the 6 base AT rich site is moved through the -7 to -16 ATG upstream area (Table 2). Of the 41,941 motifs found, 1,875 corresponded to the annotated transcription initiation site area. 12 of the 1,875 identified annotated genes were identified to contain all 9 possible 6-base AT rich potential transcriptional start site motifs.
Demonstration case 2.
Of the 1,816,638 ESTs in the Arabidopsis dataset, 36.86% of the ESTs possess multiple predicted polyA initiation recognition sites within 300 bases of the terminal 3’ base. The number of ESTs and the percentage of the polyA initiation site motif in comparison to the total number of ESTs can be found in Table 3. The recognized polyA initiation motifs ranged from a low of 6.47% to a high of 24.60% of the EST database. An increase of 36.86% of recognized motifs in comparison to the number of ESTs indicate there are ESTs with multiple polyA initiation recognition sites. While CisSERS may not be capable to differentiate which of these sites is utilized in vivo the information CisSERS provides enables an additional level of focus to which ESTs may possess multiple polyA initiation sites as well as which ESTs could be transcriptionally regulated due to premature polyA extension of the transcript.
CAPS marker development
Demonstration case 1.
To screen a population of Arabidopsis for the cfq mutation, the wild type (WT) and mutant sequences (cfq) were processed using CisSERS. The analysis revealed TaqI as an enzyme that generates clear differences due to the point mutation (Fig 2a). Biological examination of the wild type, cfq mutant, and F1 heterozygous plants was conducted through digestion of the amplified ATPC1 gene product. Visualization of the digestion pattern was completed with a 10% polyacrylamide gel (Fig 2b). The banding pattern in the biological gel matches the predicted gel image produced by CisSERS and verified the utility of this tool for an enzyme selection for CAPS analysis of this mutation. This CAPS marker is currently being deployed for screening of F1 and F2 plants and confirming the phenotypic observations demonstrating the utility of CisSERS in enabling genetics research (Cruz and Kramer, unpublished). However, Arabidopsis represents a diploid demonstration case with a very well defined genome. Such analyses can get complicated in the case of a sample with a higher ploidy as illustrated in the next demonstration case.
Demonstration case 2.
Sequencing of a selected Mdpbag (Malus x domestica putative bitter pit associate gene) gene from eight apple cultivars yielded fifteen sequences which were manually trimmed to remove plasmid and primer sequences (S4 File). Upon CisSERS evaluation, Cac8I was chosen due to its potential to differentiate five of the alleles across these eight cultivars (Fig 3A). The wet-lab gel (Fig 3B) and the predicted gel image agreed with only three of the eight digested samples. These differences likely indicate presence of additional alleles which are not expected to be represented by the draft apple genome [[20]]. However, further analysis provides a resolution to some of the differences. CisSERS predicts the restriction digest banding pattern only of the sequence used as input, and, for heterozygous organisms, the banding pattern for a restriction enzyme digest will be representative of all alleles as seen for the ‘Macintosh’ predicted digest and the actual digest (Fig 3A first 3 lanes and Fig 3B lane 1b). To resolve the latter difficulty, CisSERS has the distinctive functionality to link two or more sequences together to provide more accurate predictive gel visualization. This is demonstrated with the ‘Red Gravenstein’ samples in Fig 3A where the ‘Red_Grav4’ and ‘Red_Grav10’ alleles were linked to produce the ‘Red Grav’ composite. The linked predicted gel visualization matches the wet-lab gel from ‘Red Gravenstein’ in lane 3b of Fig 3B. The predicted gel visualization for cloned sequences from ‘Haralson’ and the actual gel demonstrates the possibility that the ‘Haralson’ cultivar is homozygous for this allele and further in-depth investigation is warranted. The wet-lab gel digests demonstrate the remaining cultivars display a combination of two alleles indicating these are heterozygous and require sequencing of additional clones to capture the other allele. Subsequent clone selection and sequencing for these cultivars captured additional alleles indicated from this analysis (data not shown). This process illustrated a case where CisSERS output identified an enzyme that was not effective at resolving all the alleles in a species with a complex genome and provided impetus to pursue additional experimentation which resulted in identification of additional alleles. Also, application of CisSERS in complex genomes with draft genome information can enable identification of areas that may have potential sequence inaccuracies.
High-Throughput Analyses
Demonstration Case.
Analysis of 78,096 cDNA sequences from Arabidopsis with 41 restriction enzymes through CisSERS produced the summary in Table 4. The individual enzymes’ restriction sites were identified in a range of 1.63% to 81.28% of the total number of sequences. BssHII restriction sites represent the least number of sites in the 78,096 cDNA dataset and the largest number of restriction sites for a single enzyme was found for Bst6I at 168,429 sites. These results demonstrate the unique capability of CisSERS to process large datasets for guiding enzyme selection decisions for global applications such as reduced representation sequencing.
Discussion
The results demonstrate the effectiveness and multiple distinctive functionalities of CisSERS in analyzing mature sequence data. Identifying enzymes for CAPS markers was highly effective in the cfq example. Interestingly, in the case of Mdpbag sequences, the resulting information about the probable heterozygosity of the locus was critical and resulted in further investigation revealing the presence of additional alleles, thus guiding further wet lab research. The highly customizable and diverse motif detection functionality resulted in the identification of potential AT rich transcriptional start sites in the Nostoc genome. The versatility of CisSERS is evident by using the motif identification feature to predict prokaryotic promoter architecture and eukaryotic poly-adenylation (polyA) initiation recognition sites. Canonical as well as non-canonical Pribnow and polyA initiation site cis-element motifs in the sequence upstream of the coding sequence areas of Nostoc sp. PCC 7107, NC_019676.1, and sequence upstream of the 3’ UTR area of Arabidopsis thaliana was searched and potential cis-elements were identified to facilitate future investigation. Lastly, the high-throughput analysis capabilities of CisSERS were demonstrated. Analyzing entire transcriptomes or genomes enables data-guided decision making for subsequent restriction enzyme based experimentation. High-throughput sequencing technologies are expensive and experimental design is a major component prior to sequencing. Based on the effective identification of restriction sites in standard and custom sequences, the identification of enzymes for reduced representation sequencing is also expected to be accurate and help ensure quality experimental design prior to sequencing. Combined, these experiments confirm the biological applicability of CisSERS as a highly effective addition to researcher’s toolkits.
Limitations and Future Improvements
CisSERS is a comprehensive and useful tool as demonstrated in previous sections. Extremely large datasets may require higher amounts of RAM or lengthy run times when processing all enzymes and gel visualization of these datasets may cause noticeable computer lag. These limitations are overcome by increasing RAM and computer processing speeds but can also be alleviated by decreasing the amount of input sequences or the number of enzymes being processed. CisSERS relies on base sequence as supplied in the fasta format which does not hold any type of sequence specific methylation data and as such any methylation susceptible restriction enzyme site identification by CisSERS would have to be scrutinized. At this point, methylation identification is simply based on any change in function of the restriction enzyme by methylation including: requiring methylation, requiring no-methylation and any partial specificity. Currently, each of these enzymes must be further evaluated by the user to make sure the chosen enzyme fits their project and the type of DNA they are processing.
Conclusions
As a tool developed to facilitate biological approaches, CisSERS enables the identification of restriction sites and custom motifs in large mature multi-sequence data files. Genotyping by sequencing and reduced representation sequencing approaches commonly utilize a restriction enzyme and CisSERS provides an efficient platform that will aid in the decision making process for users to determine the number of sites across the genome or transcriptome of interest. This is expected to facilitate guided development and deployment of CAPS markers for breeding and restriction enzyme selection for mutation identification that leverage the polymorphisms present in populations. Additionally, the custom motif functionality provides a convenient tool to query assembled genomes and transcriptome datasets for regions of biological interest. Overall, CisSERS is a standalone, open source front end tool for efficient and prudent utilization of next-generation sequencing data as the science begins to shift focus from how much data can be obtained to how we best utilize these data.
Supporting Information
S1 File. CisSERS program in .jar format available for download.
https://doi.org/10.1371/journal.pone.0152404.s001
(JAR)
S4 File. Sequences used for CAPS marker development in this study.
https://doi.org/10.1371/journal.pone.0152404.s004
(PDF)
Acknowledgments
The authors would like to thank Vandhana Krishnan, Washington State University for useful discussions.
Author Contributions
Conceived and designed the experiments: AD RMS TK. Performed the experiments: RMS TK MG MSC AH JG. Analyzed the data: RMS TK AD AK KE DMK. Contributed reagents/materials/analysis tools: AD AK KE DMK. Wrote the paper: RMS TK AD.
References
- 1. Van Tassell CP, Smith TPL, Matukumalli LK, Taylor JF, Schnabel RD, Lawley CT, et al. SNP discovery and allele frequency estimation by deep sequencing of reduced representation libraries. Nature Methods. 2008;5(3):247–52. ISI:000253777900018. pmid:18297082
- 2. Sanchez CC, Smith TPL, Wiedmann RT, Vallejo RL, Salem M, Yao JB, et al. Single nucleotide polymorphism discovery in rainbow trout by deep sequencing of a reduced representation library. Bmc Genomics. 2009;10. Artn 559 ISI:000272791700001.
- 3. Eveland AL, McCarty DR, Koch KE. Transcript profiling by 3'-untranslated region sequencing resolves expression of gene families. Plant Physiology. 2008;146(1):32–44.
- 4. Koepke T, Schaeffer S, Krishnan V, Jiwan D, Harper A, Whiting M, et al. Rapid gene-based SNP and haplotype marker development in non-model eukaryotes using 3 ' UTR sequencing. Bmc Genomics. 2012;13. Artn 18.
- 5.
Krishnan V. COMPUTATIONAL APPROACHES FOR COMPARATIVE GENOMICS AND TRANSCRIPTOMICS USING 454 SEQUENCING TECHNOLOGY: Washington State University; 2009.
- 6. Baird NA, Etter PD, Atwood TS, Currey MC, Shiver AL, Lewis ZA, et al. Rapid SNP Discovery and Genetic Mapping Using Sequenced RAD Markers. PLOS One. 2008;3(10). Artn E3376 ISI:000265121600002.
- 7. Chutimanitsakun Y, Nipper RW, Cuesta-Marcos A, Cistue L, Corey A, Filichkina T, et al. Construction and application for QTL analysis of a Restriction Site Associated DNA (RAD) linkage map in barley. Bmc Genomics. 2011;12:4.
- 8. Gaudet M, Fara AG, Sabatti M, Kuztninsky E, Mugnozza GS. Single-reaction for SNP genotyping on agarose gel by allele-specific PCR in black poplar (Populus nigra L.). Plant Molecular Biology Reporter. 2007;25(1–2):1–9.
- 9. Shu YJ, Li Y, Zhu ZL, Bai X, Cai H, Ji W, et al. SNPs discovery and CAPS marker conversion in soybean. Molecular Biology Reports. 2011;38(3):1841–6. ISI:000286941400050. pmid:20859693
- 10. Agarwal M, Shrivastava N, Padh H. Advances in molecular marker techniques and their applications in plant sciences. Plant Cell Reports. 2008;27(4):617–31. ISI:000254205200001. pmid:18246355
- 11. Neff MM, Turk E, Kalishman M. Web-based primer design for single nucleotide polymorphism analysis. Trends in Genetics. 2002;18(12):613–5. ISI:000179433300007. pmid:12446140
- 12. Ilic K, Berleth T, Provart NJ. BlastDigester—a web-based program for efficient CAPS marker design. Trends in Genetics. 2004;20(7):280–3. ISI:000222710500003. pmid:15219390
- 13. Thiel T, Kota R, Grosse I, Stein N, Graner A. SNP2CAPS: a SNP and INDEL analysis tool for CAPS marker development. Nucleic Acids Research. 2004;32(1). ARTN e5 ISI:000188988700005.
- 14. Taylor J, Provart NJ. CapsID: a web-based tool for developing parsimonious sets of CAPS molecular markers for genotyping. BMC Genetics. 2006;7. Artn 27 ISI:000237846500001.
- 15. Zhang RF, Zhu ZH, Zhu HM, Nguyen T, Yao FX, Xia K, et al. SNP Cutter: a comprehensive tool for SNPPCR-RFLP assay design. Nucleic Acids Research. 2005;33:W489–W92. ISI:000230271400099. pmid:15980518
- 16. Chang HW, Cheng YH, Chuang LY, Yang CH. SNP-RFLPing 2: an updated and integrated PCR-RFLP tool for SNP genotyping. Bmc Bioinformatics. 2010;11. Artn 173 ISI:000277004500001.
- 17. Vincze T, Posfai J, Roberts RJ. NEBcutter: a program to cleave DNA with restriction enzymes. Nucleic Acids Research. 2003;31(13):3688–91. ISI:000183832900087. pmid:12824395
- 18. Chang HW, Yang CH, Chang PL, Cheng YH, Chuang LY. SNP-RFLPing: restriction enzyme mining for SNPs in genomes. Bmc Genomics. 2006;7. Artn 30 ISI:000235715000001.
- 19.
Williams G. EMBOSS Trimest. Available: http://emboss.sourceforge.net/apps/release/6.6/emboss/apps/trimest.html2001.
- 20. Velasco R, Zharkikh A, Affourtit J, Dhingra A, Cestaro A, Kalyanaraman A, et al. The genome of the domesticated apple (Malus x domestica Borkh.). Nature Genetics. 2010;42(10):833–9. ISI:000282276600012. pmid:20802477