4) Recommended Browser (problems with Internet Explorer)

1. Preparing sequences and files for submission

1a) Alignment of the sequences

• The use of an external program (such as ClustalW from NCBI or SeqMan from DNAStar) is required to produce the correct alignment of all the sequences.

Note that preprocessing of sequences (e.g., alignment, vector removal) is not performed by SHMTool.

• If sequences are obtained from different sources (such as different animals, experiments or clonal subgroups), they can be aligned separately and saved in different FASTA multi-sequence files, so one or more files can be uploaded for each category (CONTROL or CASE). If this option is used, the same mutation found in more than one alignment will be considered as a unique event within each alignment and considered as a unique mutation. Note that you can decide how many independent alignments you want to include in each category (CONTROL or CASE) and that the number of each does not have to be the same. On the other hand, the existence or not of multiple alignments assigned to each category (CONTROL or CASE) is not relevant when non-unique (total) counts are computed. 
• We recommend incorporating the consensus sequence in every alignment, so each FASTA multi-sequence file will contain the consensus sequence and all the other sequences together.

1b) Length of the sequences

    Because the sequencing reaction often leads to sequences of differing lengths, it may be tempting in these circumstances to count mutations up to the end of each sequence, especially if mutations are rare. However, to ensure correctness in the calculation of mutation frequencies, we require each mutated sequence to span the exact length of the consensus. Otherwise, we are likely to underestimate the unique mutation frequency since not all sites would be represented by all sequences.

This does mean that any sequences shorter or longer than the consensus cannot be used.

    To achieve this:

• Deletions must be annotated with non-sequence symbols (dashes, underscores, periods, “N”, etc…) and will be ignored by the algorithm.
• Insertions must be eliminated from the sequences by deleting the additional nucleotides to assure that the length of the aligned sequences is exactly the same.
• Check that all alignments within your CONTROL and CASE groups present the same length, and that this is also true for the consensus sequence.

1c) Format of the sequences

    Sequences in FASTA formatted files are required by SHMTool. In this format, every sequence is preceded by a line starting with a " >" symbol, followed by the name of the sequence and the sequence itself (without spaces between nucleotides). For example:

>CONSENSUS

CAGGTCCAACTGCAGCAGGCAGCAGCCCTGGGGCTGAGGTTGTGAAGCCTGGTGAAGCTGTCCTGCAAGGCTTCTGGCT

ACACCTTCACCAGTCAAGAACAAGGCTGGGTGAAGCAGAGGCGCGGACGAGGCCTTGAGTGGATTGGAAGGATTGAGCC

TTACAGTGGTGATACTAAGTACAATGAGAAGTTCAAGAACAAGGCCACACTGACTGTAGACAAACCGTCCAGCACAGCC

TACATGCAGCTCAGCAGCCTGACATCTGATTATTGTGCAAGTCAAGA

>6733_D08_MS-3723475_032.ab1

CAGGTCCAACTGCAGCAGGCAGCAGCCCTGGGGCTGAGGTTGTGAAGCCTGGTGAAGCTGTCCTGCAAGGCTTCTGGCT

ACACCTTCACCAGTCAAGAACAAGGCTGGGTGAAGCAGAGGCGCGGACGAGGCCTTGAGTGGATTGGAAGGATTGAGCC

TTACAGTGGTGATACTAAGTACAATGAGAAGTTCAAGAACAAGGCCACACTGACTGTAGACAAACCGTCCAGCACAGCC

TACATGCAGCTCAGCAGCCTGACATCTGATTATTGTGCAAGTCAAGA

Note that if you are using a word-processor (such as Microsoft Word) to edit your FASTA files, these need to be saved in "Plain Text" format.

2. Uploading the files

Recommended Browser (problems with Internet Explorer)

2a) Chosing the number of files within CONTROL and CASE categories

      SHMTool provides the opportunity to segregate the sequences in subgroups when they have been obtained from different sources (such as different animals, experiments or clonal subgroups), so they can be aligned separately and saved in different FASTA multi-sequence files. Therefore, it is possible to assign several FASTA files to the same category (CONTROL or CASE) by changing the number of input files in the main page. SHMTool generates independent counts for each FASTA file, before pooling the data within each category and generating the overall comparisons of CONTROL and CASE.

       If this option is used, the same mutation found in more than one alignment will be considered as a unique event within each alignment and counted as a unique mutation. On the other hand, the existence or not of multiple alignments assigned to each category (CONTROL or CASE) is not relevant when non-unique or total counts are computed.

Note that you can decide how many independent alignments you want to include in each category (CONTROL or CASE) and that the number of each does not have to be the same for both categories.

2b) Uploading files into CONTROL and CASE categories

You may upload FASTA files from your computer into CONTROL and CASE categories.

2c) Labeling the consensus sequence

    A single consensus sequence has to be chosen for whole pair-wise comparison of CONTROL versus CASE.

• The actual sequence of the consensus must be incorporated in at least one of the FASTA multisequence files used in the analysis. We recommend incorporating the consensus sequence in every alignment, so each FASTA multisequence file will contain the consensus sequence and the number of other sequences together.

Use the exact label that designates the consensus sequence in the FASTA multi-sequence files (in the example above, this is CONSENSUS)

• To allow the algorithm to distinguish which sequence in the alignment is the consensus, the FASTA label used for the consensus sequence must be introduced in the main page.

2d) Deciding if subregion(s) will be analyzed

    SHMTool provides the option to analyze separately one or multiple subregions within the region of interest. This would be especially useful if polymorphisms, structural variants or different domains are localized within the region of study. If you localize these subregions providing their exact position or limits in the box appearing in the main page, SHMTool will generate the output results independently for every fragment of the region. Notice that this is not necessary when the whole region is going to be analyzed.

2e) Submit your pair-wise comparison

Now you are ready to run SHMTool. The results will temporally remain on the server for 48 hours to facilitate your acces to them and then will be deleted.  

3. Output tables and results

3.a) Example Datasets

We provide here one example of CONTROL and CASE datasets that can be used to test the main features of SHMTool. There are 3 CONTROL and 4 CASE files in FASTA format, which correspond to Jh2-Jh4 intron data obtained from three wild-type and four Msh6^-/- age-matched mice, respectively. These datasets were used in a previously published work (Li et al., 1996) to demonstrate that deficiency on the mismatch repair protein MSH6 is associated with an overall decrease in the mutation frequency, reduced mutations in A:T pairs and increased transitions in C:G pairs. Each file contains a consensus sequence of length 600 bp (named CONSENSUS, Genbank accesion number NT_166318.1, nucleotides 256620185-25620784) and 10-35 actual sequences, some of which may be mutated. The test of this in vivo example using SHMTool allows to confirm the key results reported previously in Li et al., 1996 (note that there are minor quantitative differences with the reported results since the more stringent preprocessing required for SHMTool was not used for the original study) and even provides a more extensive analysis of the data, validating and supporting the use of SHMTool as a webserver for comparative analysis of somatic hypermutation datasets.

As 'Consensus Sequence Label' use: CONSENSUS

3.b) Mutations Per Sequence

SHMTool represents in histograms the number of sequences (frequency in y-axis) that contain X mutations (mutations per sequence in x-axis). In the example below, the majority of the sequences are unmutated in both CONTROL and CASE categories, but many sequences contain from 1 to 20 mutations compared to the CONSENSUS.

As shown in the tables below, the raw histogram data may be downloaded in a tab-delimited version and exported to other applications for further analysis or graph representation. The first row indicates the absolute number of mutations detected in a sequence (mutPerSeq). The rest of the rows show the number of sequences, within each FASTA file, that contain the corresponding number of mutations indicated in the first row.

The actual list of mutations per sequence may be downloaded in a tab-delimited version. In the example below, 0 mutations are detected in sequences named ‘5311_F02_MS-272791_003.ab1’ from Control group or ‘S409022.AEC-232-M13F.ab1’ from Case group; 1 mutation is detected in sequence named ‘5311_G06_MS-272807_018.ab1’, and so on. Using this option we can easily detect which sequences are mutated and how many mutations they carry.

3.c) Mutation profiles

• Unique mutations: Within the files, each of which may contain many sequences, identical mutations (i.e. same mutation, same site) will be considered unique and counted only once (analysis of non-unique counts, where every sequence is considered independent, is available separately). Unique mutations are classified (by base pair from→to, e.g. C→T) and aggregated separately for CONTROL and CASE categories. Use 'download' link to download the raw data (in tab-delimited format) and export it to a different application (e.g. Excel).

MUT: type of substitution (sum: compiled mutations; Tv: transversions; Ts: transitions; GC: mutations at C or G sites; AT: mutations at A or T sites; ALL: overall mutations) CT: number of accumulated mutations in CONTROL category. All unique mutations from the different files that constitute the category are pooled together. CA: number of accumulated mutations in CASE category. All unique mutations from the different files that constitute the category are pooled together. SITES: number of potential sites in the consensus sequences that can present an specific mutation. NCT: number of CONTROL trials. We consider it correct to set NCT to the theoretical maximum number of mutations, i.e. the number of SITES multiplied by the number of groups (files) in the CONTROL category. NCA: number of CASE trials. We consider it correct to set NCA to the theoretical maximum number of mutations, i.e. the number of SITES multiplied by the number of groups (files) in the CASE category. PR_CT: mutation frequency in the CONTROL category. It includes correction by base composition since it is calculated dividing CT by NCT. PR_CA: mutation frequency in the CASE category. It includes correction by base composition since it is calculated dividing CA by NCA. P: p-value when χ2 test [as implemented by the R function prop.test (Bates et al., 1996) is applied to 2x2 contigency tables as described above. NA: is shown if minimal conditions for statistical test reliability are not met, and therefore a P value cannot be given.

This table corresponds to the example data provided in section (3.a)

• Non-unique mutations (motifs): Within the files, each of which may contain many sequences, all mutations at a specific position within a motif (which appears underlined) are scored and each sequence is considered independent. For non-unique mutation counts, all Q sequences in each category are considered independently and L=NQ (where N is the number of sites, and L is the theoretical maximum count). The actual frequency of independent mutational events is expected to lie somewhere between the unique and non-unique mutation frequencies (Rada et al., Mol Cell 2004).

• Motif positions: The position on the consensus sequence of the underlined/mutated residue from every motif may be downloaded and further explored if desired. Note that if no motifs of a specific type are found, a blank page will be returned.

This table corresponds to the example data provided in section (3.a)

---

3.e) Mutation profiles (Subregion and non-Subregion)

The user may also specify a subregion (potentially non-contiguous subset of sites) S of the consensus sequence to be analyzed separately. The complementary subregion S’ (all sites not in S) is also analyzed.

All the statistics for unique and non-unique mutations comparing S and S’ are also generated.

The subregion feature will typically be used, for example, to analyze complementary determining regions that form the antigen binding sites separately from the framework regions that position the CDRs in the variable region, or to exclude known polymorphic sites that are not mutations and would lead to an overestimate of the mutation frequency.

Remarkably, besides the basic analysis where S (or S', independently) is compared between CASE and CONTROL, SHMTool provides a detailed comparative analysis of S versus S' within each category.

Note that when performing subregion/non-subregion analysis for motifs, it is the position of the mutable base (e.g. the "C" in a WRC motif) that is considered within or outside the subregion. Thus, even if part of the motif lies outside the subregion, it will be counted as lying inside if the mutable base is within the subregion.

S and S' regions are compared within CONTROL category	S and S' regions are compared within CASE category
_SUB: refers to S region (subregion), which is defined by the user during the uploading process. _NON: refers to S' region (non-subregion), which is automatically deduced by SHMTool as the complementary to S region.	Both unique and non-unique mutations are submitted to this kind of subregion analysis.

These table corresponds to the example data provided in section (3.a) when, as an example, nucleotides 1-300 are defined as Subregion.

---

3.f) Spatial plots

• Plots along the length of the sequence: These plots show the distribution of mutations along the length of the sequence, expressed as a percentage of sequences mutated at each site. Only non-unique (total) mutations are reported here. Mutations in the CONTROL dataset are shown in the top panel, those for the CASE dataset in the bottom panel of each plot.

This plot corresponds to the example data provided in section (3.a)

• Specific graphs for each from (C/G/A/T) category: Below the first plot (that shows overall mutations) there are specific graphs for each from (C/G/A/T) category, denoted as "Mutations from C", "Mutations from G", etc. The data shown in each of these graphs is simply site-filtered from the first overall plot (the "Mutations from C" plot shows only mutations at C sites, and so on). In each category plot, we also highlight sites depending on whether they are relevant hot-, cold-spots or other (i.e. neither hot- or cold-spots). Each site is color-coded as follows.

• Mutations at C: WRC hotspot (yellow), SYC coldspot (cyan/dashed), other C-sites (gray/dotted)
• Mutations at G: GYW hotspot (yellow), GRS coldspot (cyan/dashed), other G-sites (gray/dotted)
• Mutations at A: WA hotspot (yellow), other A sites (gray/dotted)
• Mutations at T: TW hotspot (yellow), other T sites (gray/dotted)

These plots correspond to the example data provided in section (3.a)

---

4. Recommended Browser (problems with Internet Explorer)

Microsoft Internet Explorer causes two inconveniences, both associated with use of the browser "Back" button.

1. When using the "Back" button to return to the main results menu page, IE insists on a manual refresh.
2. When going further back to the entry page, one finds all data entries have disappeared and have to be re-entered.

We therefore recommend using either Mozilla Firefox or Safari web browsers.

---