5-5: The T-Score web application – Scoring and ranking peptides for MHC binding

PHP Data structure for scoring matrices

We wish to import the scoring matrix data from the .csv files in a data structure in PHP that would be convenient to use for calculating the peptide binding scores. There is indeed one structure that is perfect for the job: an associative array or dictionary, in which we have a key for each amino-acid letter that has, as associated value, an array with the scoring values as elements.

To exemplify the concept, let’s consider the lines for Alanine and Leucine in the .csv file for MHC allele HLA-A1 that we generated in the previous section:

A;1.69;1.50;-0.15;-1.5;0.82;1.36;1.00;1.14;-1.20
L;0.51;1.62;1.24;-0.29;0.19;0.44;0.38;0.22;1.31

and manually build a dictionary called $matrix with just two keys, “A” and “L”, for now.



$matrix = array(
    "A" => array(1.69,1.50,-0.15,-1.5,0.82,1.36,1.00,1.14,-1.20), 
    "L" => array(-1.00,0.00,-1.00,-1.00,0.00,-1.38,0.83,0.12,-1.67)
);

$matrix = array(

"A" => array(1.69,1.50,-0.15,-1.5,0.82,1.36,1.00,1.14,-1.20),

"L" => array(-1.00,0.00,-1.00,-1.00,0.00,-1.38,0.83,0.12,-1.67)

);

We could use this matrix to score a 9 mer peptide solely composed of Alanine and Leucine residues, with sequence: ALLAAAALA

The scoring array for an amino-acid can be accessed very easily in the matrix by using the amino-acid itself as key. For an Alanine (“A”), this would be:

$matrix[“A”]

Then, to access the score for the A in a specific position of the peptide, we just use the position itself as an array index, where position 1 in the peptide corresponds to index 0 (the first element) in the array. So the score for an alanine in position 1 would be:

$matrix[“A”][0]

Check out the following scoring example:


<?php

$matrix = array(
    "A" => array(1.69,1.50,-0.15,-1.5,0.82,1.36,1.00,1.14,-1.20), 
    "L" => array(-1.00,0.00,-1.00,-1.00,0.00,-1.38,0.83,0.12,-1.67)
);

$peptide = "ALLAAAALA";

$score = 0;

$aminoacids = str_split($peptide, 1);

$i = 0;
foreach($aminoacids as $aminoacid){
    $score = $score + $matrix[$aminoacid][$i];
    $i++;
}

echo "<p>$peptide score is $score</p>";  // ALLAAAALA score is 1.29
?>

<?php

$matrix = array(

"A" => array(1.69,1.50,-0.15,-1.5,0.82,1.36,1.00,1.14,-1.20),

"L" => array(-1.00,0.00,-1.00,-1.00,0.00,-1.38,0.83,0.12,-1.67)

);

$peptide = "ALLAAAALA";

$score = 0;

$aminoacids = str_split($peptide, 1);

$i = 0;

foreach($aminoacids as $aminoacid){

$score = $score + $matrix[$aminoacid][$i];

$i++;

}

echo "<p>$peptide score is $score</p>"; // ALLAAAALA score is 1.29

Based on this code, we can easily write a function that takes a peptide and a matrix as arguments and returns the peptide score. In the following code example we write the function and then use it to score a peptide. In the function, we include a check to make sure that each analysed amino-acid exists as a key in the matrix and return an error if it does not. This could be useful for code debugging and will prevent PHP errors if we supply an invalid sequence.


<?php

function pepscore($peptide, $matrix){
    $score = 0;
    $aminoacids = str_split($peptide, 1);
    $i = 0;
    foreach($aminoacids as $aminoacid){
        if(array_key_exists($aminoacid, $matrix)){
            $score = $score + $matrix[$aminoacid][$i];
            $i++;
        }
        else{
            return "aa not found";
        }
    }
    return $score;
}

$matrix = array(
    "A" => array(1.69,1.50,-0.15,-1.5,0.82,1.36,1.00,1.14,-1.20), 
    "L" => array(-1.00,0.00,-1.00,-1.00,0.00,-1.38,0.83,0.12,-1.67)
);
$peptide = "ALLAAAALA";
$score = pepscore($peptide, $matrix);

echo "<p>$peptide score is $score</p>";  // ALLAAAALA score is 1.29

?>

<?php

function pepscore($peptide, $matrix){

$score = 0;

$aminoacids = str_split($peptide, 1);

$i = 0;

foreach($aminoacids as $aminoacid){

if(array_key_exists($aminoacid, $matrix)){

$score = $score + $matrix[$aminoacid][$i];

$i++;

}

else{

return "aa not found";

}

return $score;

}

$matrix = array(

"A" => array(1.69,1.50,-0.15,-1.5,0.82,1.36,1.00,1.14,-1.20),

"L" => array(-1.00,0.00,-1.00,-1.00,0.00,-1.38,0.83,0.12,-1.67)

);

$peptide = "ALLAAAALA";

$score = pepscore($peptide, $matrix);

echo "<p>$peptide score is $score</p>"; // ALLAAAALA score is 1.29

Once we have a score, we can use the thresholds data available in the matrix to rank the peptide among the x% best binders.

In the previous section we did store the threshold information in the csv file in this format:

tre;6.68;5.18;4.21;3.47;2.85;2.31;1.83;1.39;0.99;0.61;

We could store this data in our matrix array in the same format as the scoring data for amino-acids:


$matrix = array(
    "A" => array(1.69,1.50,-0.15,-1.5,0.82,1.36,1.00,1.14,-1.20), 
    "L" => array(-1.00,0.00,-1.00,-1.00,0.00,-1.38,0.83,0.12,-1.67),
    "tre" => array(6.68,5.18,4.21,3.47,2.85,2.31,1.83,1.39,0.99,0.61)
);

$matrix = array(

"A" => array(1.69,1.50,-0.15,-1.5,0.82,1.36,1.00,1.14,-1.20),

"L" => array(-1.00,0.00,-1.00,-1.00,0.00,-1.38,0.83,0.12,-1.67),

"tre" => array(6.68,5.18,4.21,3.47,2.85,2.31,1.83,1.39,0.99,0.61)

);

Let’s calculate the score and the rank this time.


<?php

$matrix = array(
    "A" => array(1.69,1.50,-0.15,-1.5,0.82,1.36,1.00,1.14,-1.20), 
    "L" => array(-1.00,0.00,-1.00,-1.00,0.00,-1.38,0.83,0.12,-1.67),
    "tre" => array(6.68,5.18,4.21,3.47,2.85,2.31,1.83,1.39,0.99,0.61)
);

$peptide = "ALLAAAALA";

$score = 0;

$aminoacids = str_split($peptide, 1);

$i = 0;
foreach($aminoacids as $aminoacid){
    $score = $score + $matrix[$aminoacid][$i];
    $i++;
}

$rank = "11"; // We start by assuming the rank is higher than 10. We basically bulk assign, arbitrarily, a rank 11 to all the peptides with rank higher than 10. 
// if the score is higher than one of the thresholds, we will upgrade that.
// Mind that the lower the rank number, the best the peptide is as a binder. A peptide with rank 1 would be among the 1% of best binders.  
// This is a better rank than being "just" among the 10% best binders, so make no mistakes here.

$i = 1;
foreach($matrix["tre"] as $threshold){
    if($score >= $threshold){
        $rank = $i;
        break;
    }
    else{
        $i++;
    }
}

echo "<p>$peptide score is $score, which ranks it among the $rank% best binders</p>";  // ALLAAAALA score is 1.29, which ranks it among the 9% best binders
?>

<?php

$matrix = array(

"A" => array(1.69,1.50,-0.15,-1.5,0.82,1.36,1.00,1.14,-1.20),

"L" => array(-1.00,0.00,-1.00,-1.00,0.00,-1.38,0.83,0.12,-1.67),

"tre" => array(6.68,5.18,4.21,3.47,2.85,2.31,1.83,1.39,0.99,0.61)

);

$peptide = "ALLAAAALA";

$score = 0;

$aminoacids = str_split($peptide, 1);

$i = 0;

foreach($aminoacids as $aminoacid){

$score = $score + $matrix[$aminoacid][$i];

$i++;

}

$rank = "11"; // We start by assuming the rank is higher than 10. We basically bulk assign, arbitrarily, a rank 11 to all the peptides with rank higher than 10.

// if the score is higher than one of the thresholds, we will upgrade that.

// Mind that the lower the rank number, the best the peptide is as a binder. A peptide with rank 1 would be among the 1% of best binders.

// This is a better rank than being "just" among the 10% best binders, so make no mistakes here.

$i = 1;

foreach($matrix["tre"] as $threshold){

if($score >= $threshold){

$rank = $i;

break;

}

else{

$i++;

}

echo "<p>$peptide score is $score, which ranks it among the $rank% best binders</p>"; // ALLAAAALA score is 1.29, which ranks it among the 9% best binders

Let’s now modify the pepscore() function we wrote before to return the score, the rank, or both, depending on a third optional argument, so that we can enjoy some flexibility while using it.


<?php

function pepscore($peptide, $matrix, $mode="all"){
    $score = 0;
    
    $aminoacids = str_split($peptide, 1);
    
    $i = 0;
    foreach($aminoacids as $aminoacid){
        if(array_key_exists($aminoacid, $matrix)){
            $score = $score + $matrix[$aminoacid][$i];
            $i++;
        }
        else{
            return "aa not found";
        }
    }
    
    $rank = "11"; // We start by assuming the rank is higher than 10. We basically bulk assign, arbitrarily, a rank 11 to all the peptides with rank higher than 10. 
    // if the score is higher than one of the thresholds, we will upgrade that.
    // Mind that the lower the rank number, the best the peptide is as a binder. A peptide with rank 1 would be among the 1% of best binders.  
    // This is a better rank than being "just" among the 10% best binders, so make no mistakes here.
    // The threshods, in the array, are in decreasing order. The first threshold is the highest and scores higher that that rank the peptide among the best 1% binders
    // if the score fails to be higher than the first threshold, we check for the second threshold, and so on.
    // When we find a threshold that is lower than the calculated score, we stop iterating through the thresholds by breaking the foreach cycle with a break statement
    // if we do not find such a threshold, the rank will remain 11
    
    $i = 1; // $i is the rank value here
    foreach($matrix["tre"] as $threshold){
        if($score >= $threshold){
            $rank = $i;
            break;
        }
        else{
            $i++;
        }
    }
    
    if($mode == "all"){
        return array($score, $rank);
    }
    elseif($mode == "score"){
        return $score;
    }
    elseif($mode == "rank"){
        return $rank;
    }
}

$matrix = array(
    "A" => array(1.69,1.50,-0.15,-1.5,0.82,1.36,1.00,1.14,-1.20), 
    "L" => array(-1.00,0.00,-1.00,-1.00,0.00,-1.38,0.83,0.12,-1.67),
    "tre" => array(6.68,5.18,4.21,3.47,2.85,2.31,1.83,1.39,0.99,0.61)
);

$peptide = "ALLAAAALA";
$pepscore = pepscore($peptide, $matrix);
$score = $pepscore[0];
$rank =  $pepscore[1];

echo "<p>$peptide score is $score, which ranks it among the $rank% best binders</p>";  // ALLAAAALA score is 1.29, which ranks it among the 9% best binders
?>

<?php

function pepscore($peptide, $matrix, $mode="all"){

$score = 0;

$aminoacids = str_split($peptide, 1);

$i = 0;

foreach($aminoacids as $aminoacid){

if(array_key_exists($aminoacid, $matrix)){

$score = $score + $matrix[$aminoacid][$i];

$i++;

}

else{

return "aa not found";

}

$rank = "11"; // We start by assuming the rank is higher than 10. We basically bulk assign, arbitrarily, a rank 11 to all the peptides with rank higher than 10.

// if the score is higher than one of the thresholds, we will upgrade that.

// Mind that the lower the rank number, the best the peptide is as a binder. A peptide with rank 1 would be among the 1% of best binders.

// This is a better rank than being "just" among the 10% best binders, so make no mistakes here.

// The threshods, in the array, are in decreasing order. The first threshold is the highest and scores higher that that rank the peptide among the best 1% binders

// if the score fails to be higher than the first threshold, we check for the second threshold, and so on.

// When we find a threshold that is lower than the calculated score, we stop iterating through the thresholds by breaking the foreach cycle with a break statement

// if we do not find such a threshold, the rank will remain 11

$i = 1; // $i is the rank value here

foreach($matrix["tre"] as $threshold){

if($score >= $threshold){

$rank = $i;

break;

}

else{

$i++;

}

if($mode == "all"){

return array($score, $rank);

}

elseif($mode == "score"){

return $score;

}

elseif($mode == "rank"){

return $rank;

}

$matrix = array(

"A" => array(1.69,1.50,-0.15,-1.5,0.82,1.36,1.00,1.14,-1.20),

"L" => array(-1.00,0.00,-1.00,-1.00,0.00,-1.38,0.83,0.12,-1.67),

"tre" => array(6.68,5.18,4.21,3.47,2.85,2.31,1.83,1.39,0.99,0.61)

);

$peptide = "ALLAAAALA";

$pepscore = pepscore($peptide, $matrix);

$score = $pepscore[0];

$rank = $pepscore[1];

echo "<p>$peptide score is $score, which ranks it among the $rank% best binders</p>"; // ALLAAAALA score is 1.29, which ranks it among the 9% best binders

We now have a function, pepscore(), to score and rank a peptide based on a matrix. A quite valuable and central piece of code for our T-Score application.

Let’s now see how we can generate a proper matrix associative array starting from one of the .csv files we produced in the previous section.

Generating scoring matrices from the .csv files

We will now write a little function matrix_file_to_array() to convert our .csv files into associative arrays with the format discussed above.

The flow of the code is as follows:

– We import the file to a variable
– Split on “\n” to get the lines
– For each line, we split the line to an array by using the semicolon as splitting element. The first element of this array will become a key in our final matrix associative array, while all the other elements will be part of an array assigned as value to this key. Remember how each line is structured:

A;1.69;1.50;-0.15;-1.5;0.82;1.36;1.00;1.14;-1.20

– The function then returns this matrix array


<?php

function matrix_file_to_array($matrix_file_path){
    $matrix_file = file_get_contents($matrix_file_path); // Importing the .csv file to a variable
    $lines = explode("\n", $matrix_file); // Splitting the file into lines
    $matrix_array = array(); // Our final matrix associative array, empty for now

    foreach($lines as $line){
        $line = trim($line);
        if($line != ""){ // If the line is not an empty line
            $elements = explode(";", $line); // splitting the line on the semicolons
            $key = array_shift($elements); // array_shift() removes the first element from an array and returns it. The first element is our key for the matrix associative array.
            $matrix_array[$key] = $elements; // 
        }   
    }
    
    return $matrix_array;
}

// Let's test this function on the .csv file with the data for the HLA-A1 allele

$a1_matrix = matrix_file_to_array("matrix/a1-matrix.csv");
var_dump($a1_matrix);

?>

<?php

function matrix_file_to_array($matrix_file_path){

$matrix_file = file_get_contents($matrix_file_path); // Importing the .csv file to a variable

$lines = explode("\n", $matrix_file); // Splitting the file into lines

$matrix_array = array(); // Our final matrix associative array, empty for now

foreach($lines as $line){

$line = trim($line);

if($line != ""){ // If the line is not an empty line

$elements = explode(";", $line); // splitting the line on the semicolons

$key = array_shift($elements); // array_shift() removes the first element from an array and returns it. The first element is our key for the matrix associative array.

$matrix_array[$key] = $elements; //

}

return $matrix_array;

}

// Let's test this function on the .csv file with the data for the HLA-A1 allele

$a1_matrix = matrix_file_to_array("matrix/a1-matrix.csv");

var_dump($a1_matrix);

On executing the code above from our script.php file located in the tscore directory, we get the var_dump of the matrix for the HLA-A1 allele. It will show that the $a1_matrix variable is now an array of 22 elements, as expected, as we have 20 amino-acids, plus the X amino-acid, plus the “tre” line with the thresholds.

Here is how the var_dump starts (it’s quite long so we just report the beginning of the output, but please test it on your own server):

array(22) {
[“A”]=>
array(9) {
[0]=>
string(4) “1.69”
[1]=>
string(4) “1.50”
[2]=>
string(5) “-0.15”
[3]=>
string(5) “-1.50”
[4]=>
string(4) “0.82”
[5]=>
string(4) “1.36”
[6]=>
string(4) “1.00”
[7]=>
string(4) “1.14”
[8]=>
string(5) “-1.20”
}
[“C”]=>
array(9) {
[0]=>
string(4) “2.00”
[1]=>
string(4) “0.00”
[2]=>
string(4) “0.00”
[3]=>
string(4) “1.00”
[4]=>
string(4) “0.00”
[5]=>
string(5) “-2.00”
[6]=>
string(4) “0.00”
[7]=>
string(5) “-2.00”
[8]=>
string(4) “0.00”
}
[“D”]=>
array(9) {
[0]=>
string(5) “-1.50”
[1]=>
string(5) “-2.00”
[2]=>
string(4) “6.52”
[3]=>
string(5) “-0.67”
[4]=>
string(5) “-2.00”
[5]=>
string(4) “0.67”
[6]=>
string(5) “-1.33”
[7]=>
string(5) “-1.00”
[8]=>
string(5) “-2.00”
}
etc…

Getting real: scoring and ranking a peptide sequence by using a full scoring matrix

In the following code example we will select an allele name, generate the scoring matrix array “on the fly” and use it to score and rank an actual peptide sequence.

To get a fully woking self-contained example, we move the pepscore() and matrix_file_to_array() functions to a “functions.php” file created in a folder called “include”, inside the tscore directory.

We also create a new php file called “generate_matrices.php” where we will move all the code related to the download of the .html files from the original website and the generation of the .csv files in our matrix folder, from these .html files. The reason for moving some of the code in a different file is to keep our main file, script.php, as clean and light as possible so that it will be more readable and easier to maintain. As functions.php, we will store this file in the include folder.

Our application file structure therefore becomes, for now:

tscore (permission 777)
    matrix (permission 777)
    script.php
    include
        functions.php
        generate_matrices.php

And here is the content of the files:

generate_matrices.php


<?php
$urls = array("http://www.imtech.res.in/raghava/nhlapred/matrices/a1.html", "http://www.imtech.res.in/raghava/nhlapred/matrices/a2.html", "http://www.imtech.res.in/raghava/nhlapred/matrices/a3.html");
 
foreach($urls as $url){
    preg_match("/a\d\.html/", $url, $matches); // We are looking for the final "ax.html" part of the URL
    $page_name = $matches[0]; // and store it into a variable called $page_name
    $local_page_path = "matrix/$page_name";  // We will store the files with their original names in the matrix folder, here we define the relative path with respect to script.php
    if(!file_exists($local_page_path)){  // We check that the file is not there already, and if not
        file_put_contents($local_page_path, file_get_contents($url)) or die("something went wrong");
    }
}

$file_names = scandir("matrix"); // An array with all the names of the files in the matrix directory
 
$score_regexp = "/<TD><CENTER>(.{1,5})<\/CENTER><\/TD>/";
$threshold_regexp = "/<td><center>(.{4})<\/center><\/td>/"; // Unlike those for the score values, html tags for the threshold values are in lowercase in the original pages we are scraping the data from
 
foreach($file_names as $file_name){ // Iterating through the files, one by one
    if(preg_match("/^(.+)\.html$/", $file_name, $matches) and !file_exists("matrix/".$matches[1]."-matrix.csv")){
        $html_file_path = "matrix/$file_name";
        $csv_file_path = "matrix/".$matches[1]."-matrix.csv";
        $html_file_contents = file_get_contents($html_file_path);
        $html_file_lines = explode("\n", $html_file_contents);
        $csv_handle = fopen($csv_file_path, "w");
        $table_flag = 0; // Each time we come across an opening table tag during file parsing, this will be raised by one. 
        $threshold_csv_line_temp = "\ntre;";
        $j = 1;  // Counter for the thresholds lines
        // File parsing to scrape the data may begin!
        
        foreach($html_file_lines as $line){
            $line = trim($line); // You never know which weird characters may be lurking at the beginning or end of your lines...
            
            if(preg_match("/<table/", $line)){
                $table_flag++;
            }
            
            if($table_flag == 5){ // We are parsing one of the lines belonging the scores table
                if(preg_match_all($score_regexp, $line, $matches) and !preg_match("/P1/", $line)){
                    $i = 1; 
                    foreach($matches[1] as $match){
                        if($i < 10){
                            fwrite($csv_handle, $match.";"); // We want to add a semicolon after each of the element but the last one and we have a total of 10 elements, one aminoacid letter plus 9 scores
                            $i++;
                        }
                        elseif($i == 10){
                            fwrite($csv_handle, $match."\n"); // After the last score we add a newline instead of a semicolon
                        }
                    } 
                }
            }
            
            elseif($table_flag == 6){ // We are parsing one of the lines belonging the thresholds table
                if(preg_match($threshold_regexp, $line, $matches)){  // We use preg_match instead of preg_match_all as in the original files HTML source each threshold is in it's own line so we just need to grab one value here
                    // we also manage things a little differently, adding the values to a variable instead of writing them to the file directly, for the same reason.
                    if($j < 10){
                        $threshold_csv_line_temp = $threshold_csv_line_temp.$matches[1].";";
                        $j++;
                    }
                    elseif($j == 10){ // The last threshold value
                        $threshold_csv_line_temp = $threshold_csv_line_temp.$matches[1];
                    }
                }
            }
            
        }
        
        fwrite($csv_handle, $threshold_csv_line_temp);
        fclose($csv_handle);     
 
    }
}
?>

<?php

$urls = array("http://www.imtech.res.in/raghava/nhlapred/matrices/a1.html", "http://www.imtech.res.in/raghava/nhlapred/matrices/a2.html", "http://www.imtech.res.in/raghava/nhlapred/matrices/a3.html");

foreach($urls as $url){

preg_match("/a\d\.html/", $url, $matches); // We are looking for the final "ax.html" part of the URL

$page_name = $matches[0]; // and store it into a variable called $page_name

$local_page_path = "matrix/$page_name"; // We will store the files with their original names in the matrix folder, here we define the relative path with respect to script.php

if(!file_exists($local_page_path)){ // We check that the file is not there already, and if not

file_put_contents($local_page_path, file_get_contents($url)) or die("something went wrong");

}

$file_names = scandir("matrix"); // An array with all the names of the files in the matrix directory

$score_regexp = "/<TD><CENTER>(.{1,5})<\/CENTER><\/TD>/";

$threshold_regexp = "/<td><center>(.{4})<\/center><\/td>/"; // Unlike those for the score values, html tags for the threshold values are in lowercase in the original pages we are scraping the data from

foreach($file_names as $file_name){ // Iterating through the files, one by one

if(preg_match("/^(.+)\.html$/", $file_name, $matches) and !file_exists("matrix/".$matches[1]."-matrix.csv")){

$html_file_path = "matrix/$file_name";

$csv_file_path = "matrix/".$matches[1]."-matrix.csv";

$html_file_contents = file_get_contents($html_file_path);

$html_file_lines = explode("\n", $html_file_contents);

$csv_handle = fopen($csv_file_path, "w");

$table_flag = 0; // Each time we come across an opening table tag during file parsing, this will be raised by one.

$threshold_csv_line_temp = "\ntre;";

$j = 1; // Counter for the thresholds lines

// File parsing to scrape the data may begin!

foreach($html_file_lines as $line){

$line = trim($line); // You never know which weird characters may be lurking at the beginning or end of your lines...

if(preg_match("/<table/", $line)){

$table_flag++;

}

if($table_flag == 5){ // We are parsing one of the lines belonging the scores table

if(preg_match_all($score_regexp, $line, $matches) and !preg_match("/P1/", $line)){

$i = 1;

foreach($matches[1] as $match){

if($i < 10){

fwrite($csv_handle, $match.";"); // We want to add a semicolon after each of the element but the last one and we have a total of 10 elements, one aminoacid letter plus 9 scores

$i++;

}

elseif($i == 10){

fwrite($csv_handle, $match."\n"); // After the last score we add a newline instead of a semicolon

}

elseif($table_flag == 6){ // We are parsing one of the lines belonging the thresholds table

if(preg_match($threshold_regexp, $line, $matches)){ // We use preg_match instead of preg_match_all as in the original files HTML source each threshold is in it's own line so we just need to grab one value here

// we also manage things a little differently, adding the values to a variable instead of writing them to the file directly, for the same reason.

if($j < 10){

$threshold_csv_line_temp = $threshold_csv_line_temp.$matches[1].";";

$j++;

}

elseif($j == 10){ // The last threshold value

$threshold_csv_line_temp = $threshold_csv_line_temp.$matches[1];

}

fwrite($csv_handle, $threshold_csv_line_temp);

fclose($csv_handle);

}

functions.php


<?php

function pepscore($peptide, $matrix, $mode="all"){
    $score = 0;
    
    $aminoacids = str_split($peptide, 1);
    
    $i = 0;
    foreach($aminoacids as $aminoacid){
        if(array_key_exists($aminoacid, $matrix)){
            $score = $score + $matrix[$aminoacid][$i];
            $i++;
        }
        else{
            return "aa not found";
        }
    }
    
    $rank = "11"; // We start by assuming the rank is higher than 10. We basically bulk assign, arbitrarily, a rank 11 to all the peptides with rank higher than 10. 
    // if the score is higher than one of the thresholds, we will upgrade that.
    // Mind that the lower the rank number, the best the peptide is as a binder. A peptide with rank 1 would be among the 1% of best binders.  
    // This is a better rank than being "just" among the 10% best binders, so make no mistakes here.
    // The threshods, in the array, are in decreasing order. The first threshold is the highest and scores higher that that rank the peptide among the best 1% binders
    // if the score fails to be higher than the first threshold, we check for the second threshold, and so on.
    // When we find a threshold that is lower than the calculated score, we stop iterating through the thresholds by breaking the foreach cycle with a break statement
    // if we do not find such a threshold, the rank will remain 11
    
    $i = 1; // $i is the rank value here
    foreach($matrix["tre"] as $threshold){
        if($score >= $threshold){
            $rank = $i;
            break;
        }
        else{
            $i++;
        }
    }
    
    if($mode == "all"){
        return array($score, $rank);
    }
    elseif($mode == "score"){
        return $score;
    }
    elseif($mode == "rank"){
        return $rank;
    }
}

function matrix_file_to_array($matrix_file_path){
    $matrix_file = file_get_contents($matrix_file_path); // Importing the .csv file to a variable
    $lines = explode("\n", $matrix_file); // Splitting the file into lines
    $matrix_array = array(); // Our final matrix associative array, empty for now

    foreach($lines as $line){
        $line = trim($line);
        if($line != ""){ // If the line is not an empty line
            $elements = explode(";", $line); // splitting the line on the semicolons
            $key = array_shift($elements); // array_shift() removes the first element from an array and returns it. The first element is our key for the matrix associative array.
            $matrix_array[$key] = $elements; // 
        }   
    }
    
    return $matrix_array;
}
?>