Skip to content
Advertisement

Split fasta file into specific new fasta files

So I’m writing this code that will read a fasta file. In the fasta file, there will be 10 sequences. The start of the sequence will be “>” I want to split 50:50 of those sequences and create two new fasta files with it. 5 sequences in one new file; the other 5 sequences in another new file.

I have an interface that is the sequence itself, and the fasta file class with 10 sequences implements it. I have two classes for the first 5 sequences and the last 5 sequences. I have a reader that reads the fasta file. I just don’t know how to write something that will split it into two files. Do I add another class that splits and then another one that creates? I’m unsure how to go about this. please help ty

Example fasta file format:

>XP_012508841.1 PREDICTED: apolipoprotein E [Propithecus coquereli]
---------------------------MNALWAVVVVTFLAGCQAKVEPALE------LE
AGELTEWQTGQPWEQALGRFWDYLRWVQTLSNQVQEELLSSQVTQELTVLMEETMKEVKA
YKLELEEQMSPMAEETRARLSKELQAAQARLGQDMQDLRGRLEQYCSEAQAVLGQNTEEL
RARLAAHLRKLRKRLLRDAEDLQKRLAVYQAGAREGAERGVSAIRERLRPLVERGRTRAA
TVG-TLASQPLRERAQAWGERLRAGLEEMGARGRDRLDEVREQLEEVRAKVEEQAAQLRL
QAEAFESRLKSWFTPLVEDMQRQWAGLVEKLQAAVQSPTATPEPSTNH

>XP_012616817.1 apolipoprotein E [Microcebus murinus]
---------------------------MNALWAVVVVTFLAGCQAKVEPALE------PE
ARELTEWQTGQPWEQALGRFWDYLRWVQTLSDQVQEELLSSQVTQELTVLMEETMKELKA
YKSELEQQLNPMAEETRARLSKELQAAQARLGQDMQDLRGRLAQYRGEVQAMLGQSTDEL
RARFASHLRKLRKRLLRDAEDLQRRLAVYQAGAREGAERGVSAIRERLGPLVEQGRLRAA
TVG-TLAGQPLRERAEAWGERLRARLEEMGSQGRDRLDEVREQVEEVRVKVEEQAAKMRL
QAEALEARLKSWFTPLVEDMQRQWAGLVEKLQAAVQGTTATPEATENH

>XP_008047463.1 apolipoprotein E [Carlito syrichta]
---------------------------MKVLWAVLLVTFLAGGRADVEPELE------TE
LRKPAEWQSGQPWEAALGRLWDYLRWVQTLSDEVQEELLSSQVTQELT------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------------------
------------------------------------------------

Advertisement

Answer

I initially created a method (function) to parse out Sequences from a Fasta File but like so many other things Fasta files can be in so many formats that some form of setup will need to be done to parse one type from another type. Your Fasta files are no exception, since in most Fasta sequences I don’t believe the hyphen or minus character (-) is allowed within a sequence unless it is contained within the sequence Header but, I can see that in your Fasta files, hyphens are contained within sequences. Of course, I am most likely wrong about that. 😉

With this now known I’ve added some additional variables that can be manually set so as to provide more flexibility towards different Fasta file formats (I hope). In reality this should actually be a class instead of a method right from the beginning but I’m going to let you convert it to a workable class youself.

Now, this is a large method (and I don’t feel good about that) with a lot of comments but I wanted to give you something relatively quick. I recommend you read all the provided comments within the code.

After you try this method…do make a class of it:

/**
 * Returns a {@code List<String>} Interface object of all the Sequence Clusters 
 * detected within the supplied Fasta Data File.<br>
 * 
 * @param sourceFilePath (String) The full path and file name to the source 
 * Fasta data file to parse.<br>
 * 
 * @param destinationFileFolderPath (String)  The full path to the
 * destination directory (folder) where the two split Fasta files are to be 
 * created. If null or Null String ("") is supplied then the two split Fasta 
 * files will be created within the source file directory. Destination split 
 * file names are auto-generated!<br>
 * 
 * @param splitRatio (String) The ratio of sequence clusters to be applied to 
 * each created split Fasta files. Ratio is considered percentages and the 
 * highest percentage is first followed by the lower percentage delimited with 
 * a colon (:), for example: "60:40". The two values provided <b>must</b> sum 
 * to 100.<br>
 * 
 * @param newSequenceDesignator (String) By default an inner blank line is 
 * considered the end of a sequence cluster and the possible start of a new 
 * one. If a designator is supplied here then blank lines are ignored within 
 * the source file while parsing (if they exist). If there are no blank lines 
 * within the source file separating the sequence clusters then a designator 
 * <b>must</b> be supplied, usually the Sequence Cluster Header designator (>) 
 * is used. The New Sequence Designator can be the same as the Sequence Cluster 
 * Header Designator.<br>
 * 
 * @return ({@code List<String>}) The List of parsed Sequence Clusters from 
 * the supplied source Fasta data file.
 */
public static java.util.List<String> splitFastaFile(String sourceFilePath, String destinationFileFolderPath, 
                                 String splitRatio, String newSequenceDesignator) {
    /* System newline character(s) to use for console 
       display when required.   */
    String ls = System.lineSeparator();
    
    // Valid characters allowed to be contained within a Fasta sequence line.
    String allowableCharactersInSequences = "ABCDEFGHIJKLMNOPQRSTUVWXYZ-*";
    
    /* By default, any sequence cluster line (other than the Header) which 
       contains an invalid character (a character not stipulated within the
       'allowableCharactersInSequences' variable) is ignored and not added 
       to the Sequence Cluster. If however, you provide a string character 
       or phrase to the 'replaceInvalidSequenceCharactersWith' variable then 
       all invalid characters will be replace by what is held within that
       variable. Keep in mind that you should use a character or case phrase 
       that is not considered valid (where as it is not also contained within  
       the 'allowableCharactersInSequences' variable). You want to always 
       maintain an invalid sequence line as INVALID unless the replacement is 
       indeed a VALID responce (repair) to the sequence which makes up the 
       sequence line or Sequence Cluster.             */
    String replaceInvalidSequenceCharactersWith = null;     // Default is null.
    
    /* The String or string character that will denote 
       the start of a new Sequence cluster.  */
    String sequenceDesignatorString = ">";
    
    /* The String or string character that will denote
       the start of a Sequence Header line. Can be the 
       same as the Sequence Designator String.   */
    String sequenceHeaderDesignatorString = ">";
    
    // Escape RegEx meta characters in allowableCharactersInSequences (if any).
    allowableCharactersInSequences = allowableCharactersInSequences.replaceAll("[\W]", "\\$0");
    
    // Add the sequence Header (if any)
    boolean keepSequenceHeader = true;
    
    // Add a blank line between sequence clusters in created Split Fasta files.
    boolean blankLineBetweenSequenceClusters = false;
    
    /* If a comment is supplied to this variable then it MUST start with a 
       semicolon (;). If it doesn't then it will be nulled. This comment 
       will be applied as the fist line of any Split File created. The comment 
       provided can utilize one or all three of the method tags available.
       These tags are:        
            %%H      The High percentabe Fasta sequences split value.
            %%L      The Low percentabe Fasta sequences split value.
            %%SV     The Split Value currently being processed.   
            %%SFN    The Source File Name
    
       An example might be as what was provided below:     */
    String splitFileComment = ";SPLIT FASTA FILE - Source File: %%SFN - Percent of source: %%SV%%";
    
    // See if the supplied source Fasta file exists.
    File f = new File(sourceFilePath);
    if (!f.exists() || !f.canRead()) {
        System.err.println("splitFastaFile() method error! Either the "
                + "specified source file can not be found or permission "
                + "to read the file does not exist!");
        return null;
    }
    
    /* Get the supplied Fasta file name. The destination
       files will derive from this name.        */
    String sourceFileName = f.getName();

    /* If null or null string ("") is passed as the destination 
       directory then the Source file directory will also become 
       the destination for the two created Split Files.         */  
    if (destinationFileFolderPath == null || destinationFileFolderPath.isEmpty()) {
        String absPath = new File(sourceFilePath).getAbsolutePath();
        destinationFileFolderPath = absPath.substring(0, absPath.lastIndexOf(File.separator));
    }
    
    /* Make sure the supplied destination file folder path
       contains a system file separator character ( or /).   */
    if (!destinationFileFolderPath.endsWith(File.separator)) {
        destinationFileFolderPath = destinationFileFolderPath + File.separator;
    }

    /* Make sure a proper Split Ratio is supplied! The Split Ratio is
       supplied as a colon (:) delimited string containg the desired 
       percentage of sequences to be saved within the first text file
       and the desired percentage of sequences to be saved to the second
       text file. If you want 80%% of the sequences within the Fasta source
       file to be written to the first text file and you want the remaining
       20%% to be written to the second text file then you would supply to
       the 'splitRatio' parameter:  "80:20". Whatever is supplied, the sum
       of the two supplied values MUST equal 100 (100%%). The higher of the 
       two values MUST be first (this is enforced).              */
    if (!splitRatio.matches("\d{1,3}:\d{1,3}")) {
        System.err.println("splitFastaFile() method Error! An invalid Split "
                + "Ratio string was supplied! (" + splitRatio + ") The format "
                + "must be: "80:20".");
        return null;
    }
    /* Split the ratio string provided within the 'splitRatio' parameter.
       Convert the string numerical values to Integer and check validity. */
    String[] ratioParts = splitRatio.split("\s*:\s*");
    int high = Integer.valueOf(ratioParts[0]);
    int low = Integer.valueOf(ratioParts[1]);
    if (high + low != 100) {
        System.err.println("splitFastaFile() method Error! An invalid Split "
                + "Ratio string was supplied! (" + splitRatio + ") The percentage "
                + "values supplied must sum to 100. What was supplied sums to: "
                + (high + low) + ".");
        return null;
    }
    else if (high < low) {
        System.err.println("splitFastaFile() method Error! An invalid Split "
                + "Ratio string was supplied! (" + splitRatio + ") The higher "
                + "percentage value must be on the left and the lower percentage "
                + "value on the right! Swapping values to make it valid (" 
                + low + ":" + high + ")!");
        int tmp = high; 
        high = low; 
        low = tmp;
    }

    // Load source file Sequences into a List Interface object.
    List<String> sequenceList = new ArrayList<>();
    int fileLineCount = 0;
    int validDataLineCount = 0;
    int clusterLineCount = 0;
    try (BufferedReader reader = new BufferedReader(new FileReader(sourceFilePath))) {
        StringBuilder sb = new StringBuilder("");
        String line;
        while((line = reader.readLine()) != null) {
            /* Increment the File Lines Count (tracks the 
               total number of lines in Fasta source file). */
            fileLineCount++;
            /* Trim off leading and trailing whitespaces, tabs, 
               etc from the read in data file line.   */
            line = line.trim();
            // Ignore lines that START with a simicolon (;). They are considered comment lines.
            if (line.startsWith(";")) { continue; }
            // Ignore first line of file if it's blank.
            if (line.isEmpty() && validDataLineCount == 0) { continue; }
            // See if this is a Sequence Header Line and whether or not we are to keep it.
            if (line.startsWith(sequenceHeaderDesignatorString) && !keepSequenceHeader) { 
                // Increment the Sequence Cluster Line Count.
                clusterLineCount++;
                continue; // Ignore this Header...loop again.
            }
            if (validDataLineCount > 0 && ((sequenceDesignatorString != null && !sequenceDesignatorString.isEmpty()) ? line.startsWith(sequenceDesignatorString) : line.isEmpty())) {
                String tmpLine = "";
                if (line.startsWith(sequenceDesignatorString)) { 
                    tmpLine = line; 
                }
                sequenceList.add(sb.toString());
                sb.delete(0, sb.length());
                clusterLineCount = 0;
                if (!tmpLine.isEmpty()) {
                    sb.append(tmpLine).append(ls);
                    clusterLineCount++;
                    validDataLineCount++;
                }
            }
            else {
                /* Skip blank lines if it is set to not be the new Sequence designator.
                   By default, blank inner file lines are considered the designator for
                   the end of a Sequence Cluster and the beginning of a new one. if 
                   however the sequenceDesignatorString variable contains a designator 
                   then blank lines will be ignored.                    */
                if ((sequenceDesignatorString != null && !sequenceDesignatorString.isEmpty()) && line.isEmpty()) {
                    continue;
                }
                // Check sequence line validity
                else if (!line.startsWith(sequenceHeaderDesignatorString) && !line.matches("[" + allowableCharactersInSequences + "]+")) {
                    // NOT VALID!
                    // Do invalid Characters get replaced...
                    if (replaceInvalidSequenceCharactersWith != null && 
                                    !replaceInvalidSequenceCharactersWith.isEmpty()) {
                        StringBuilder tmpSB = new StringBuilder("");
                        for (int i = 0; i < line.length(); i++) {
                            String s = line.substring(i, i + 1);
                            if (s.matches("[" + allowableCharactersInSequences + "]+")) {
                                tmpSB.append(s);
                            }
                            else {
                                tmpSB.append(replaceInvalidSequenceCharactersWith);
                            }
                        }
                        clusterLineCount++;
                        sb.append(tmpSB.toString()).append(ls);
                        validDataLineCount++;
                        continue;
                    }
                    // Display the Invalid sequence line detected in console and...
                    System.err.println("Invalid sequence line character(s) detected in line #" + clusterLineCount + " of sequence cluster #" + (sequenceList.size() + 1) + ", on file "
                            + "line #" + fileLineCount + "!" + ls + line);
                    // point out the characters that are invalid with Caret (^)
                    // characters under the invalid file line.
                    for (int i = 0; i < line.length(); i++) {
                        System.out.print(!line.substring(i, i + 1)
                                .matches("[" + allowableCharactersInSequences + "]+") ? "^" : " ");
                    }
                    System.out.println();
                }
                else {
                    clusterLineCount++;
                    sb.append(line).append(ls);
                    validDataLineCount++;
                }
            }
        }
        if (!sb.toString().isEmpty()) {
            sequenceList.add(sb.toString());
        }
    } catch (FileNotFoundException ex) {
        System.err.println(ex);
        return null;
    } catch (IOException ex) {
        System.err.println(ex);
        return null;
    }

    // Below is used for testing the method code...
    //System.out.println("There are " + sequenceList.size() + " sequence(s) within the source file.");
    //for (String str : sequenceList) {
    //    System.out.println(str);
    //}

    /* Auto-Generate destination file names. Destination file names will be in
       the format of:
                      "Split_{source File Name}_{High Percentage}.txt"
            and       "Split_{source File Name}_{Low Percentage}.txt"

       If the source file name is: "FastaFile010.txt" and the Split Ratio 
       supplied is: "60:40"  then the destination file names will be:
                      "Split_FastaFile010_60.txt"
            and       "Split_FastaFile010_40.txt"
    */
    String name = sourceFileName.substring(0, sourceFileName.lastIndexOf("."));
    String extension = sourceFileName.substring(sourceFileName.lastIndexOf("."));
    String highFileName = "Split_" + name + "_" + String.valueOf(high) + extension;
    String lowFileName = "Split_" + name + "_" + String.valueOf(low) + extension;

    //System.out.println(highFileName);  // For Testing...
    //System.out.println(lowFileName);   // For Testing...

    /* Determine the number of sequences for each file
       based on the number sequences contained within
       the Sequence List and the supplied desired Ratio. */
    int highSeqs = (int)((sequenceList.size() * high) * 0.01d);
    int lowSeqs = sequenceList.size() - highSeqs;
    
    // System.out.println(highSeqs + " | " + lowSeqs);     // For Testing...

    // Create the two Split Files based on the desired ratio...
    int c = 0;
    String destPath;
    int alreadyWritten = 0;
    // Loop to write create or overwrite two seperate text files.
    while (c < 2) {
        destPath = destinationFileFolderPath + (c == 0 ? highFileName : lowFileName);
        String uFormat = "UTF-8";       // Save text files in UTF-8 format.
        try {
            FileOutputStream outputStream = new FileOutputStream(destPath);
            OutputStreamWriter outputStreamWriter = new OutputStreamWriter(outputStream, uFormat);
            try (BufferedWriter bufferedWriter = new BufferedWriter(outputStreamWriter)) {
                // Apply a comment (if there is one) to the start of the current Split File.
                if (splitFileComment != null && !splitFileComment.isEmpty()) {
                    // Replace method tags (if any - %%H, %%L, %%SV):
                    String actFileComment = splitFileComment.trim()
                                            .replace("%%H", String.valueOf(high))
                                            .replace("%%L", String.valueOf(low))
                                            .replace("%%SV", (c == 0 ? String.valueOf(high) : String.valueOf(low)))
                                            .replace("%%SFN", sourceFileName);
                    bufferedWriter.write(actFileComment);
                    bufferedWriter.write(System.lineSeparator());
                }
                int i;
                for (i = alreadyWritten; i < sequenceList.size(); i++) {
                    bufferedWriter.write(sequenceList.get(i));
                    /* Add a blank line after a sequence cluster is written to 
                       file in preparation for the next sequence but don't do 
                       this if it's the last sequence in the required set to be 
                       written. If you don't want a blank line between all your 
                       sequence clusters in the saved files then supply false to 
                       the 'blankLineBetweenSequenceClusters' variable.    */
                    if (blankLineBetweenSequenceClusters && 
                            (i + 1 - alreadyWritten) != (c == 0 ? highSeqs : lowSeqs)) {
                        bufferedWriter.write(System.lineSeparator());
                    }
                    // Write everything in buffer to file right away.
                    bufferedWriter.flush();
                    /* Stop writing to the first file is we've reach our
                       sequence limit for this file. Break out of the loop
                       so that file #2 can be written.           */
                    if (c == 0 && (i + 1) == highSeqs) { break; }
                }
                // Update how many sequences we've already written to files.
                alreadyWritten = i+1;
                // Close the current writter.
                bufferedWriter.close();
            }
        }
        catch (IOException e) {
            System.err.println(e);
            break;
        }
        /* Increment for which file we are currently writing to:
           0 = file #1 and 1 = file #2.         */
        c++;
    }
    return sequenceList;
}

Usage example:

List<String> seqs = splitFastaFile("MyFastaFile.txt", null, "60:40", null);
if (seqs.size() > 0) {
    System.out.println("There were " + seqs.size() + " Sequence Clusters within the Fasta source file!");
}
else {
    System.err.println("There was Error while processing the supplied Fasta File!");
}

User contributions licensed under: CC BY-SA
3 People found this is helpful
Advertisement