Human MLPA Probe Design (H-MAPD): a probe design tool for both electrophoresis-based and bead-coupled human multiplex ligation-dependent probe amplification assays

Multiplex ligation-dependent probe amplification (MLPA) has proven to be an efficient and reliable technique for gene dosage analysis. This technique was first introduced in 2002 [1], and has been widely applied in a variety of clinical and research situations. In this method, two sequence-tagged half probes are annealed to adjacent sites on the genomic target sequence and ligated using a thermostable DNA ligase. The ligated probes are subsequently amplified with universal PCR primers (one of which is fluorescently labeled) and quantified using electrophoresis (each product has a distinct size, which allows for identification) (See additional file 1: Diagram of electrophoresis-base MLPA). A typical, capillary-based MLPA assay allows for the quantification of up to 45 distinct sequences. In a more recent development, FlexMAP bead-coupled MLPA has been described, which allows for the simultaneous detection of 100 distinct sequences per reaction (identification of distinct sequences, which can all be the same size, is based on association with a distinct bead) [2] (See additional file 2: Diagram of bead-coupled MLPA).

Compared to other gene dosage detection methods, including, for example, Southern blot analysis and FISH (fluorescent in situ hybridization), MLPA is high throughput, requires only small amounts of starting DNA (cf. Southern blotting), does not require cells for chromosome spreads (as in FISH) and can be used to target any genomic sequences for copy number analysis, irrespective of their size or proximity to each other. Quantitative PCR (qPCR) is a potential alternative, but its use in a multiplex assay is limited by the spectral overlap of the fluorescent dyes used [1], and detecting 2:1 or 3:2 copy number change is challenging.

However, while kits covering multiple regions/genes of interest are available from MRC-Holland, for those regions not yet commercially available as kits, there are significant challenges in designing custom assays for MLPA. In the original description, MLPA probe sets were generated by cloning the target specific sequence into a family of M13 derived vectors that already contained the variable length fragments. With improvements in oligonucleotide synthesis, completely synthetic probe sets have been successfully used for MLPA reactions, obviating the need for the implementation of the M13 method [3]. The current major obstacle to successful MLPA probe design is the long list of criteria that need to be simultaneously satisfied to improve the likelihood of a successful assay. These criteria include probe length, Tm, secondary structure, GC content, nucleotide composition at the ligation site, sequence uniqueness, avoidance of known SNPs, etc. Manual probe design is time consuming and error prone. Our software automates this tedious MLPA probe design process.

Our software applies to human genomic MLPA probe design only but will be generalized in the future to deal with other genomes. In addition, H-MAPD is currently only for the design of MLPA copy number assays, but may be generalized in the future to deal with the design of MLPA assays for other applications, such as SNP detection and methylation changes. For multiple input sequences, the stuffer and bead tag modifications will be applied to the input sequences sequentially. For example, in bead-coupled MLPA, the first bead tag will be inserted into the first input sequence and so on.

The software takes into consideration most factors predicted to affect MLPA probe performance. However, there are likely to be unexpected factors that might cause problems. These factors include secondary structure of genomic DNA at the hybridization site, the presence of un-described SNP(s) at the hybridization site, sequence errors in the reference genome assembly, etc. The software performs sequence homology searches using BLAT due to its speed and excellent programming interface. Another popular sequence search tool is BLAST. Because they are implemented using different algorithms, BLAT and BLAST may not return identical results. The authors strongly recommend that users verify H-MAPD results with BLAST for independent validation.

Two platforms are available for MLPA assays. The electrophoresis-based platform requires the use of capillary electrophoresis systems which are available in most institutions; however long oligonucleotides (> 150 nt) in high quality are difficult to synthesize currently. The bead-coupled platform works with half probes (LPO/RPO) less than 100 nucleotides in length, but requires the use of a Luminex system. For the electrophoresis-based platform, probe length increase is achieved either by inserting stuffer sequences between the PCR primer and hybridizing sequences, or by extending the length of hybridizing sequences. By using stuffer sequences, hybridizing sequences have identical length and therefore have a more unified Tm. Since the hybridizing sequences are short, the use of stuffer provides the potential for detecting mutations/polymorphisms that are in close proximity to each other. However, designing stuffer sequences takes time. The stuffer sequences in additional file 3 have been verified in advance so that the union of default PCR primer (used by the commercial MRC-Holland MLPA kits) and stuffer sequences (default left primer + left stuffer; right stuffer + default right primer) are free of secondary structure or significant homology in the human genome. Some workers prefer extending hybridizing sequences to the use of stuffer sequences. However, longer hybridizing sequences are not suitable for short target regions and tend to result in more non-specific binding to the target genome. For the bead-coupled platform, the tag sequence can theoretically be incorporated in either the LPO or RPO. Currently, since anti-tags are linked at their 5' end to the commercially available FlexMAP beads, we recommend inserting the tag sequences in the RPO as illustrated in additional file 2. Thus, only the right PCR primer is in physical proximity to the bead, minimizing any steric hindrance between the PCR product and the bead. Another reason favouring inserting the tag in the RPO is that quite a few tag sequences start with adenosine (additional file 4). These tags will affect probe signal strength if inserted between the left primer and LHS. The commercial FlexMAP tag sequences were not designed specifically for MLPA assays, and some tag sequences, when attached to the default right primer, will form significant secondary structure on their own. For example, the tags corresponding to bead 062 and 071, when attached to the default right primer, have secondary structures that are significant: ΔG = -1.538 and -1.514 respectively (additional file 4). To avoid the use of these tags, the user can insert a short dummy sequence (for example, ACGT) in the input sequence corresponding to that tag (for example, when designing an assay with tags attached to the right primer, input sequence 62 should be input as a dummy sequence, so that tag 062 is consumed by the dummy sequence). Future development of H-MAPD should allow users to use their own stuffer sequences or bead tag sequences. H-MAPD allows users to specify custom PCR primers. However, users should be careful that the union of custom primers to the stuffer/tag sequences may result in secondary structures or significant homology to the human genome, even before specific hybridizing sequences are appended.

It is a challenge to design MLPA probes that can distinguish closely related sequences in the genome. For highly similar but non-identical sequences, H-MAPD will treat them as non-specific matches and is likely to fail the probe set. One can design MLPA probes for highly related sequences by allowing multiple perfect matches to amplify the common identical fragments. Of course the result represents multiple targets, and can not be used to distinguish the highly related but different sequences. A practical solution, as indicated in the MRC-Holland probe design guidelines [4], is to find the exact difference between the sequence of interest and its related sequences, and place the difference at the end of LPO or RPO. A branch of H-MAPD specifically designed for this purpose will be implemented in the future.

H-MAPD is a web-based MLPA probe design program implemented in Perl-CGI, and hosted on a Linux server. With the increasing application of MLPA in biomedical research and the lack of free probe design software, we hope that H-MAPD will become a valuable tool for automatic MLPA probe generation and selection.

H-MAPD is freely available to non-commercial users at URL http://genomics01.arcan.stonybrook.edu/mlpa/cgi-bin/mlpa.cgi. Commercial users should contact the authors. Since the software utilizes the UCSC genome browser and UNAFold, commercial users also need to obtain a license for those programs.