MorphoCatcher: a multiple-alignment based web tool for target selection and designing taxon-specific primers in the loop-mediated isothermal amplification method

Implementation

MorphoCatcher is implemented in Python 3.0 programming language and has a web interface. The algorithm is based on the counting of all taxon-specific mutations, excluding strains-specific polymorphisms, that corresponds to each taxon (or group) in multi-species and multi-strain MSA. Then, a sliding-window function is used to visualize the polymorphism density of potential target for each taxon of the MSA. Thus, MorphoCatcher accelerates the target selection process and allows to select the target region with a high density of taxon-specific mutations.

Sliding-window approach

To calculate the density of taxon-specific mutations, the following formula of so-called average mutation index was used: ${\displaystyle {\overline{S}}_f=\frac{1}{w}\sum _{i=0}^{w-1}{M}_{f-i}=\frac{M_f+M_{f-1}+\cdots +M_{f-w+2}+M_{f-w+1}}{w},}$ where ${\overline{S}}_f$ is the simple moving average (SMA) value for a fragment (f) of MSA with a length of 20 nucleotides, which corresponds to the primer annealing site length; w—the sliding window size; and M_f−i—the number of mutations in the fragment (f) (M_f−i ∈ N₀).

The obtained value of SMA refers to the central fragment of the selected sliding window, therefore, we define the amplicon length formed by an odd number of fragments. Thus, we anchor the value ${\overline{S}}_f$ to the central fragment of sliding window, not to the abstract MSA locus.

From its previous value, the SMA can be obtained using the following recurrence formula: ${\displaystyle {\overline{S}}_f={\overline{S}}_{f-1}-\frac{M_{f-w}}{w}+\frac{M_f}{w},}$ where ${\overline{S}}_f$ is the SMA value for a fragment ( f) of MSA; ${\overline{S}}_{f-1}$—the previous SMA value; M_f−w—the number of mutations in the first fragment; M_f—the number of mutations in the current sliding window after its shift by one fragment to the right.

Web interface

MorphoCatcher uses a free “Agency” HTML5 page template from the “Start Bootstrap” service (https://startbootstrap.com) and supports the current releases of Chrome, Firefox, and Safari browsers. Computation time required for analysis usually takes few seconds depending on the server workload. The MorphoCatcher web service is freely available only for non-commercial and academic users at http://morphocatcher.ru. All output files can be downloaded as a ZIP-archive with the unique ID of the project. Graphic files have a publication quality and are available in different formats (*.eps, *.pdf, *.png, *.svg, and *.tiff).

Command line version

A Python script of the MorphoCatcher tool can be used in the Linux terminal for large-scale research projects. A command line Python script of the tool is available at the GitHub public repository: https://github.com/shrshkv/MorphoCatcher.

Manual and tutorial files

A detailed manual and tutorial files for two hypothetical case studies for the MorphoCatcher were developed to show the main applications and functions of the plugin tool during specific primer design. The case studies explain a simple bioinformatics protocol that can be used for the following example tasks: (i) to detect the bacterial potato pathogens from the Dickeya genus (Czajkowski et al., 2015) and (ii) distinguish two serotypes of potato virus Y (PVY) (Karasev & Gray, 2013) by the LAMP method. The manual demonstrates required file formatting and shows how to perform successful primer design using other well-known online bioinformatics services. Some recommendations were suggested for specific primer design using the alignment of orthologous target gene or whole-genome sequences. The manual and tutorial files are also available at https://github.com/shrshkv/MorphoCatcher.

Primer design

After the alignment processing using MorphoCatcher, an output file from the “Input for PrimerExplorer” column that corresponds to the taxon of interest can be directly submitted to the PrimerExplorer 5.0 web server (Tomita et al., 2008; Eiken Chemical Co. LTD, Tokyo, Japan; http://primerexplorer.jp/e).

The recommended MorphoCatcher protocol

Step 1: defining closely related biological taxa. First, the user should to collect several strains of the target species (or any other taxon) and the most closely related sequences from other species. The main idea of this step is to form groups of strains which should be distinguished by the LAMP diagnostics. These groups can be represented by different taxonomic ranks (e.g., species, genera, families, etc.), serotypes, or other groups of organisms. Total number of strains are usually limited by performance of the alignment tool used.

Step 2: selecting target nucleotide sequences. At this step, the user needs a set of potential targets in the taxon of interest, which can be obtained using the following services: ASAP (Glasner et al., 2006), Insignia (Phillippy et al., 2009), Panseq (Laing et al., 2010), and EDGAR (Blom et al., 2016). Optimal length of an amplicon for LAMP is about 130–200 bp, but sequences up to 300 bp are also acceptable (Notomi et al., 2000). Nevertheless, to obtain a wide variety of primer sets, we recommend the target size more than 1,000 bp. We also recommend to extract target sequence from the type strain with complete genome status.

Step 3: exporting homologous sequences. To provide statistically reliable taxon-specific mutations, different sequences should be extracted not only from the taxon of interest, but also from other closely related taxa. Related orthologous nucleotide sequences can be exported from the NCBI Nucleotide database (https://www.ncbi.nlm.nih.gov/nuccore), NCBI GenBank annotation of genome, or NCBI BLAST service (https://blast.ncbi.nlm.nih.gov) (Camacho et al., 2009).

To export orthologs using the NCBI BLAST service, a target nucleotide sequence should be used as the reference. The set of orthologous sequences can be saved from the NCBI BLAST output as a single FASTA text file. Then, the FASTA-header should be modified in accordance with two alternative formats:

>[Taxon ID]—[Target gene ID]—[Strain ID]

or

>[Taxon ID]—[Strain ID]

For example, the user can insert instead of [Taxon ID], [Target gene ID], or [Strain ID] any text that will be associated with organisms of interest.

To facilitate the rapid sequence searching on large datasets, a number of high-performance tools for multicore microprocessors can also be used, including PLAST (Nguyen & Lavenier, 2009), BLAST+ (Camacho et al., 2009), and DCBLAST (Yim & Cushman, 2017). The tools require some skills in command line of Linux computer systems.

Note, that if the taxon of interest belongs to insufficiently studied organisms and there are no orthologous sequences in the NCBI database, probably, the strategy of unique target search is more applicable for LAMP primer design.

Step 4: making multiple sequence alignment. To perform multi-species MSA of the orthologs, we recommend using the Clustal Omega service (https://www.ebi.ac.uk/Tools/msa/clustalo) (Sievers et al., 2011; Chojnacki et al., 2017). It is necessary to select an “Input” value of the “Order” menu to save the order of species in the input file. This point is important, because the MorphoCatcher algorithm recognizes several strains as belonging to the same species only if the strains have identical taxon ID in FASTA-header. The alignment should be saved as a text file.

As has been described in the previous step, the alignment step can also be accelerated by a local version of Clustal Omega algorithm and high-performance computing systems. The local run of Clustal Omega tool allows to overcome the web version limitations on input sequence number (up to 4,000 sequences or a maximum file size of 4 MB).

Step 5: masking taxon-specific mutations. We suggest using the MorphoCatcher service (http://morphocatcher.ru) for processing of the MSA. The MorphoCatcher allows the user to paste or upload the alignment input file and select the output format, which depends on the operating system (e.g., Windows or UNIX). The MorphoCatcher web interface is shown in Fig. 2.

When the alignment was uploaded, the user should click the “Catch” button which starts the processing of data to detect any taxon-specific mutations. For ease of polymorphism evaluation, we suggest the average mutation index that can be visualized as a sliding-window plot (Fig. 3). Regions with high density of taxon-specific mutations usually have a high value of the average mutation index. The peaks with maximal average mutation index and corresponding sequence can be selected as suitable locus for designing of specific primers. Then, if the target selected will be compatible with the design algorithm of PrimerExplorer, the most taxon-specific mutations will be included at the primer ends for better reaction specificity.

The most effective LAMP reactions can be performed with amplicons less than 300 bp (Notomi et al., 2000). Therefore, we use a sliding-window size of 300 bp in our algorithm. To obtain better primer specificity, we recommend to select nucleotide sequences with the average mutation index values from 2 and higher.

Once the MorphoCatcher execution of the alignment analysis is complete, the results tab appears with three columns that contain the following useful files:

• Processed MSA with highlighted species-specific mutations;

• Sliding-window plot of mutation density for each species of the alignment;

• Ready-to-use files for specific primer design using the PrimerExplorer service.

All abovementioned files can be downloaded from the MorphoCatcher output tab as a ZIP-archive. The user can also inspect positions of the species-specific mutations manually by using a processed alignment from the output files. The processed alignment file contains information about the number of species-specific mutations for each 20 nucleotides.

Step 6: designing specific primers. If the average mutation index is suitable for the taxon of interest, then the user can submit a corresponding text file from the “Input for PrimerExplorer” column directly to the PrimerExplorer 5.0 service (http://primerexplorer.jp/e). This file contains a target nucleotide sequence, conserved common region that is highlighted by asterisks (*), and detected taxon-specific mutations that highlighted by hyphen (-) symbols. These symbols mimic the PrimerExplorer format that was proposed to select any region of interest in the uploaded nucleotide sequence.

To design primer sets with a high density of taxon-specific mutations at 5′- or 3′-ends, the user should set the following parameters of the PrimerExplorer software:

• To select “Specific” mode of the “Primer option” menu;

• To click on the “Detail settings” button to turn on expert mode;

• To select a required GC-content of primers in the “Parameter condition” menu;

• To click on the “Generate” button;

• To set “Descending order” of primer sets in the “Sorting rule” menu;

• To click on the “Display” button.

After that, the user will be redirected to the page with various combinations of primer annealing sites, which can cover a different number of taxon-specific mutations (Fig. 4).

The MorphoCatcher protocol allows to solve a well-known problem of the PrimerExplorer caused by the limitation of the target sequence size up to 2,000 bp. This is possible due to our tool using an output file from the Clustal Omega algorithm that has no significant limit to the input sequence length. Thus, the user can evaluate the polymorphism level of target sequence preliminary and after that to select a suitable region (up to 2,000 bp) with high density of taxon-specific mutations for primer design.

So, after the last step of protocol the user can to select some primer sets for the oligo synthesis, reaction optimization, and assay validation basing on specificity and sensitivity tests. We recommend screening multiple primer sets in vitro, because at present there is no other effective way to predict which primer set will be recognized as a very successful primer design.

Future development

However, the MorphoCatcher also has some disadvantages. The protocol of primer design using the tool is not yet a real pipeline, therefore, we plan to improve the MorphoCatcher web service by modules for high-performance sequence similarity search and multiple alignment. We also plan to reduce the input from the user to a target sequence of interest in FASTA-format. It is will automatize further steps in according with the bioinformatics requirements to representativeness and reproducibility. We believe that these enhancements will lead to increasing of the user-friendliness and reduce complexity of the specific primer design in LAMP experiments.