'fieldset',
'#title' => t('Enter Query Sequence'),
'#description' => t('Enter one or more queries in the top text box or use the browse button to upload a file from your local disk. The file may contain a single sequence or a list of sequences. In both cases, the data must be in FASTA format. More information.. '),
'#collapsible' => TRUE,
'#collapsed' => FALSE,
);
$form['query']['FASTA'] = array(
'#type' => 'textarea',
'#title' => t('Enter FASTA sequence(s)'),
'#description'=>t('Enter query sequence(s) in the text area.'),
);
// Upload a file as an alternative to enter a query sequence
$form['#attributes']['enctype'] = 'multipart/form-data';
$form['query']['UPLOAD'] = array(
'#prefix' => 'Or upload your query files: ',
'#type' => 'file',
'#description' => t('The file should be a plain-text FASTA file and not a .doc, .docx, etc. It cannot be greater than 10 Mb in size.'),
);
// BLAST DATABASE
//.........................
$form['DB'] = array(
'#type' => 'fieldset',
'#title' => t('Choose Search Set'),
'#description' => t('Choose from one of the protein BLAST databases listed below. You can also use the browse button to upload a file from your local disk. The file may contain a single sequence or a list of sequences. '),
'#collapsible' => TRUE,
'#collapsed' => FALSE,
);
$options = get_blast_database_options('p');
$form['DB']['SELECT_DB'] = array(
'#type' => 'select',
'#title' => t('Protein BLAST Databases:'),
'#options' => $options,
'#default_value' => t('Select a database'),
);
// Upload a file as an alternative to enter a query sequence
$form['#attributes']['enctype'] = 'multipart/form-data';
$form['DB']['DBUPLOAD'] = array(
'#prefix' => 'Or upload your own dataset: ',
'#type' => 'file',
'#description' => t('The file should be a plain-text FASTA file and not a .doc, .docx, etc. It cannot be greater than 10 Mb in size.'),
);
// ALGORITHM PARAMETERS
//.........................
$form['ALG'] = array(
'#type' => 'fieldset',
'#title' => t('Algorithm parameters'),
'#collapsible' => TRUE,
'#collapsed' => TRUE,
);
//General parameters
$form['ALG']['GParam'] = array(
'#type' => 'fieldset',
'#title' => t('General parameters'),
'#collapsible' => FALSE,
);
$form['ALG']['GParam']['maxTarget'] = array(
'#type' => 'select',
'#title' => t('Max target sequences:'),
'#options' => array(
0 => t('10'),
1 => t('50'),
2 => t('100'),
3 => t('250'),
4 => t('500'),
5 => t('1000'),
6 => t('5000'),
7 => t('10000'),
8 => t('20000'),
),
'#default_value' => 2,
'#description' => t('Select the maximum number of aligned sequences to display'),
);
$form['ALG']['GParam']['shortQueries'] = array(
'#type' => 'checkbox',
'#title' => t('Automatically adjust parameters for short input sequences'),
'#default_value' => TRUE,
);
$form['ALG']['GParam']['eVal'] = array(
'#type' => 'textfield',
'#title' => t('e-value(Expect threshold)'),
'#default_value' => 10,
'#size' => 12,
'#maxlength' => 20,
'#description' => t('Expected number of chance matches in a random model.'),
);
$form['ALG']['GParam']['wordSize'] = array(
'#type' => 'select',
'#title' => t('Word size:'),
'#options' => array(
0 => t('2'),
1 => t('3'),
),
'#default_value' => 1,
'#description' => t('The length of the seed that initiates an alignment'),
);
$form['ALG']['GParam']['qRange'] = array(
'#type' => 'textfield',
'#title' => t('Max matches in a query range'),
'#default_value' => 0,
'#size' => 12,
'#maxlength' => 20,
'#description' => t('Limit the number of matches to a query range. This option is useful if many strong matches to one part of a query may prevent BLAST from presenting weaker matches to another part of the query.'),
);
// Scoring parameters
$form['ALG']['SParam'] = array(
'#type' => 'fieldset',
'#title' => t('Scoring parameters'),
'#collapsible' => FALSE,
);
$options_first = _ajax_example_get_first_dropdown_options();
$selected = isset($form_state['values']['MATRIX'] ) ? $form_state['values']['MATRIX'] : key($options_first);
$form['ALG']['SParam']['MATRIX'] = array(
'#type' => 'select',
'#title' => 'Matrix',
'#options' => $options_first,
'#default_value' => $selected,
'#description' => t('Assigns a score for aligning pairs of residues, and determines overall alignment score..'),
'#ajax' => array(
'callback' => 'ajax_example_dependent_dropdown_callback',
'wrapper' => 'dropdown-second-replace',
),
);
$form['ALG']['SParam']['gapCost'] = array(
'#type' => 'select',
'#title' => t('Gap Costs:'),
'#prefix' => '',
'#suffix' => '
',
'#options' => _ajax_example_get_second_dropdown_options($selected),
'#default_value' => 2,
'#description' => t('Cost to create and extend a gap in an alignment.'),
);
$form['ALG']['SParam']['M&MScores'] = array(
'#type' => 'select',
'#title' => t('Match/Mismatch Scores:'),
'#options' => array(
0 => t('No adjustment'),
1 => t('Composition-based statistics'),
2 => t('Conditional compositional score matrix adjustment'),
3 => t('Universal composition score matrix adjustment '),
),
'#default_value' => 2,
'#description' => t('Matrix adjustment method to compensate for amino acid composition of sequences'),
);
//Submit
$form['submit'] = array(
'#type' => 'submit',
'#default_value' => ' BLAST ',
);
return $form;
}
/**
* Form validation handler for blast_protein_form().
*
* @see blast_protein_form_validate()
*/
function blast_protein_form_validate($form, &$form_state) {
// Validate query sequence
$fastaSeq = $form_state['input']['FASTA'];
if (isset($fastaSeq)) {
if (validateFasta($fastaSeq)){
form_set_error('pBLAST', t('Error: Failed to read the Blast query: Wrong format provided for FASTA protein sequence'));
}
else {
$form_state['qFlag'] = 'seqQuery';
}
}
// Validate Query Upload
$upQuery = file_save_upload('UPLOAD', array('file_validate_extensions' => array('txt fasta fa fna')), FILE_EXISTS_RENAME);
if ($upQuery) {
$upQuery_uri = $upQuery->uri;
$form_state['upQuery_path'] = drupal_realpath($upQuery_uri);
$upQuery_content = file_get_contents($form_state['upQuery_path']);
if(validateFasta($upQuery_content)){
form_set_error('pBLAST', t('Error: Failed to upload the Blast query: Wrong format provided for FASTA protein sequence'));
}
else {
$form_state['qFlag'] = 'upQuery';
}
}
// Validate uploaded database
$upDB = file_save_upload('DBUPLOAD', array('file_validate_extensions' => array('txt fasta fa fna')), FILE_EXISTS_RENAME);
if ($upDB) {
$upDB_uri = $upDB->uri;
$form_state['upDB_path'] = drupal_realpath($upDB_uri);
$upDB_content = file_get_contents($form_state['upDB_path']);
if(validateFasta($upDB_content)){
form_set_error('DB', t('Error: Failed to upload the Blast subject sequence file: Wrong format provided for FASTA protein sequence'));
}
else {
$form_state['dbFlag'] = 'upQuery';
}
}
else {
$form_state['dbFlag'] = 'blastdb';
}
}
/**
* Form submition handler for blast_protein_form().
*
* @see blast_protein_form_submit()
*/
function blast_protein_form_submit($form, &$form_state) {
$eVal = $form_state['values']['eVal'];
$trgtKey = $form_state['values']['maxTarget'];
$numAlign = $form['ALG']['GParam']['maxTarget']['#options'][$trgtKey];
$wsKey = $form_state['values']['wordSize'];
$wordSize = $form['ALG']['GParam']['wordSize']['#options'][$wsKey];
// Expand Gap Cost key into open and extend penalties
$gapKey = $form_state['values']['MATRIX'];
switch ($gapKey) {
case 0:
$matrix ="PAM30";
$gapKey = $form_state['values']['gapCost'];
switch ($gapKey) {
case 0:
$gapOpen = 7;
$gapExtend = 2;
break;
case 1:
$gapOpen = 6;
$gapExtend = 2;
break;
case 2:
$gapOpen = 5;
$gapExtend = 2;
break;
case 3:
$gapOpen = 10;
$gapExtend = 1;
break;
case 4:
$gapOpen = 9;
$gapExtend = 1;
break;
case 5:
$gapOpen = 8;
$gapExtend = 1;
break;
}
break;
case 1:
$matrix ="PAM70";
$gapKey = $form_state['values']['gapCost'];
switch ($gapKey) {
case 0:
$gapOpen = 8;
$gapExtend = 2;
break;
case 1:
$gapOpen = 7;
$gapExtend = 2;
break;
case 2:
$gapOpen = 6;
$gapExtend = 2;
break;
case 3:
$gapOpen = 11;
$gapExtend = 1;
break;
case 4:
$gapOpen = 10;
$gapExtend = 1;
break;
case 5:
$gapOpen = 9;
$gapExtend = 1;
break;
}
break;
case 2:
$matrix ="PAM250";
$gapKey = $form_state['values']['gapCost'];
switch ($gapKey) {
case 0:
$gapOpen = 15;
$gapExtend = 3;
break;
case 1:
$gapOpen = 14;
$gapExtend = 3;
break;
case 2:
$gapOpen = 13;
$gapExtend = 3;
break;
case 3:
$gapOpen = 12;
$gapExtend = 3;
break;
case 4:
$gapOpen = 11;
$gapExtend = 3;
break;
case 5:
$gapOpen = 17;
$gapExtend = 2;
break;
case 6:
$gapOpen = 16;
$gapExtend = 2;
break;
case 7:
$gapOpen = 15;
$gapExtend = 2;
break;
case 8:
$gapOpen = 14;
$gapExtend = 2;
break;
case 9:
$gapOpen = 13;
$gapExtend = 2;
break;
case 10:
$gapOpen = 21;
$gapExtend = 1;
break;
case 11:
$gapOpen = 20;
$gapExtend = 1;
break;
case 12:
$gapOpen = 19;
$gapExtend = 1;
break;
case 13:
$gapOpen = 18;
$gapExtend = 1;
break;
case 14:
$gapOpen = 17;
$gapExtend = 1;
break;
}
break;
case 3:
$matrix ="BLOSUM80";
$gapKey = $form_state['values']['gapCost'];
switch ($gapKey) {
case 0:
$gapOpen = 8;
$gapExtend = 2;
break;
case 1:
$gapOpen = 7;
$gapExtend = 2;
break;
case 2:
$gapOpen = 6;
$gapExtend = 2;
break;
case 3:
$gapOpen = 11;
$gapExtend = 1;
break;
case 4:
$gapOpen = 10;
$gapExtend = 1;
break;
case 5:
$gapOpen = 9;
$gapExtend = 1;
break;
}
break;
case 4:
$matrix ="BLOSUM62";
$gapKey = $form_state['values']['gapCost'];
switch ($gapKey) {
case 0:
$gapOpen = 11;
$gapExtend = 2;
break;
case 1:
$gapOpen = 10;
$gapExtend = 2;
break;
case 2:
$gapOpen = 9;
$gapExtend = 2;
break;
case 3:
$gapOpen = 8;
$gapExtend = 2;
break;
case 4:
$gapOpen = 7;
$gapExtend = 2;
break;
case 5:
$gapOpen = 6;
$gapExtend = 2;
break;
case 6:
$gapOpen = 13;
$gapExtend = 1;
break;
case 7:
$gapOpen = 12;
$gapExtend = 1;
break;
case 8:
$gapOpen = 11;
$gapExtend = 1;
break;
case 9:
$gapOpen = 10;
$gapExtend = 1;
break;
case 10:
$gapOpen = 9;
$gapExtend = 1;
break;
}
break;
case 5:
$matrix ="BLOSUM45";
$gapKey = $form_state['values']['gapCost'];
switch ($gapKey) {
case 0:
$gapOpen = 13;
$gapExtend = 3;
break;
case 1:
$gapOpen = 12;
$gapExtend = 3;
break;
case 2:
$gapOpen = 11;
$gapExtend = 3;
break;
case 3:
$gapOpen = 10;
$gapExtend = 3;
break;
case 4:
$gapOpen = 15;
$gapExtend = 2;
break;
case 5:
$gapOpen = 14;
$gapExtend = 2;
break;
case 6:
$gapOpen = 13;
$gapExtend = 2;
break;
case 7:
$gapOpen = 12;
$gapExtend = 2;
break;
case 8:
$gapOpen = 19;
$gapExtend = 1;
break;
case 9:
$gapOpen = 18;
$gapExtend = 1;
break;
case 10:
$gapOpen = 17;
$gapExtend = 1;
break;
case 11:
$gapOpen = 16;
$gapExtend = 1;
break;
}
break;
case 6:
$matrix ="BLOSUM50";
$gapKey = $form_state['values']['gapCost'];
switch ($gapKey) {
case 0:
$gapOpen = 13;
$gapExtend = 3;
break;
case 1:
$gapOpen = 12;
$gapExtend = 3;
break;
case 2:
$gapOpen = 11;
$gapExtend = 3;
break;
case 3:
$gapOpen = 10;
$gapExtend = 3;
break;
case 4:
$gapOpen = 9;
$gapExtend = 3;
break;
case 5:
$gapOpen = 16;
$gapExtend = 2;
break;
case 6:
$gapOpen = 15;
$gapExtend = 2;
break;
case 7:
$gapOpen = 14;
$gapExtend = 2;
break;
case 8:
$gapOpen = 13;
$gapExtend = 2;
break;
case 9:
$gapOpen = 12;
$gapExtend = 2;
break;
case 10:
$gapOpen = 19;
$gapExtend = 1;
break;
case 11:
$gapOpen = 18;
$gapExtend = 1;
break;
case 12:
$gapOpen = 17;
$gapExtend = 1;
break;
case 13:
$gapOpen = 16;
$gapExtend = 1;
break;
case 14:
$gapOpen = 15;
$gapExtend = 1;
break;
}
break;
case 7:
$matrix ="BLOSUM90";
$gapKey = $form_state['values']['gapCost'];
switch ($gapKey) {
case 0:
$gapOpen = 9;
$gapExtend = 2;
break;
case 1:
$gapOpen = 8;
$gapExtend = 2;
break;
case 2:
$gapOpen = 7;
$gapExtend = 2;
break;
case 3:
$gapOpen = 6;
$gapExtend = 2;
break;
case 4:
$gapOpen = 11;
$gapExtend = 1;
break;
case 5:
$gapOpen = 10;
$gapExtend = 1;
break;
case 6:
$gapOpen = 9;
$gapExtend = 1;
break;
}
break;
}
// If the query was submitted via the texrfield then create a file containing it
if ( isset($form_state['qFlag']) ) {
if ( $form_state['qFlag'] == 'seqQuery' ) {
$seq_content = $form_state['values']['FASTA'];
$query = "/tmp/user__query_file.fasta";
file_put_contents ( $query , $seq_content);
}
elseif ( $form_state['qFlag'] == 'upQuery' ) {
$query = $form_state['upQuery_path'];
}
}
// If the BLAST database was uploaded then use it to run the BLAST
if ($form_state['dbFlag'] == 'upQuery') {
// Since we only support using the -db flag (not -subject) we need to create a
// blast database for the FASTA uploaded.
// NOTE: We can't support subject because we need to generate the ASN.1+ format
// to provide multiple download type options from the same BLAST
$blastdb_with_path = $form_state['upDB_path'];
system("makeblastdb -in $blastdb_with_path -dbtype prot -parse_seqids");
}
// Otherwise, we are using one of the website provided BLAST databases so form the
// BLAST command accordingly
elseif ($form_state['dbFlag'] == 'blastdb') {
$selected_db = $form_state['values']['SELECT_DB'];
$blastdb_node = node_load($selected_db);
$blastdb_with_path = $blastdb_node->db_path;
}
// Actually submit the BLAST Tripal Job
// NOTE: Tripal jobs needs to be executed from the command-line before it will be run!!
$blastdb_with_suffix = $blastdb_with_path . '.psq';
if (is_readable($blastdb_with_suffix)) {
global $user;
$output_filestub = date('YMd_His');
$job_args = array(
'program' => 'blastp',
'query' => $query,
'database' => $blastdb_with_path,
'output_filename' => $output_filestub,
'options' => array(
'evalue' => $eVal,
'word_size' => $wordSize,
'gapopen' => $gapOpen,
'gapextend' => $gapExtend,
'matrix' => $matrix
)
);
$job_id = tripal_add_job("BLAST (blastp): $query",'blast_job','run_BLAST_tripal_job', $job_args, $user->uid);
// Redirect to the BLAST results page
drupal_goto("blast/report/$job_id");
}
else {
$dbfile_uploaded_msg = ($form_state['dbFlag'] == 'upQuery') ? 'The BLAST database was submitted via user upload.' : 'Existing BLAST Database was chosen';
tripal_report_error(
'blast_ui',
TRIPAL_ERROR,
"BLAST database %db unaccessible. $dbfile_uploaded_msg",
array('%db' => $blastdb_with_path)
);
drupal_set_message('BLAST database unaccessible. Please contact the site administrator.','error');
}
}
/**
* FASTA validating parser
*
* @param $sequence
* A string of characters to be validated. A sequence in FASTA format begins with a single-line description, followed by lines of sequence data.
* The description line is distinguished from the sequence data by a greater-than (">") symbol in the first column.
* The word following the ">" symbol is the identifier of the sequence, and the rest of the line is the description (both are optional).
* There should be no space between the ">" and the first letter of the identifier. The sequence ends if another line starting with a ">" appears;
* this indicates the start of another sequence.
*
* @return
* Return a boolean. 1 if the sequence does not pass the format valifation stage and 0 otherwise.
*
*/
function validateFasta($sequence) {
$fastaIdRegEx = '/^>.*(\\n|\\r)/';
$fastaSeqRegEx = '/[^acgturykmswbdhvnxACGTURYKMSWBDHVNX\*\-\n\r]/';
if ( preg_match($fastaSeqRegEx,$sequence) && !(preg_match($fastaIdRegEx,$sequence)) ) {
$flag = 1;
} else {
$flag = 0;
}
return $flag;
}
/**
* Fill the first dropdown list with appropriate options
*
* @return
* An array consisting of matrices name for the first dropdown list
*/
function _ajax_example_get_first_dropdown_options() {
return drupal_map_assoc(array(
t('PAM30'),
t('PAM70'),
t('PAM250'),
t('BLOSUM80'),
t('BLOSUM62'),
t('BLOSUM45'),
t('BLOSUM50'),
t('BLOSUM90'),
));
}
/**
* Fill the second dropdown list with appropriate options
*
* @return
* An array containing open and extension gap values for the chosen matrix (to fill the second dropdown list)
*/
function _ajax_example_get_second_dropdown_options($key = '') {
$options = array(
t('PAM30') => drupal_map_assoc(array(
t('Existence: 7 Extension: 2'),
t('Existence: 6 Extension: 2'),
t('Existence: 5 Extension: 2'),
t('Existence: 10 Extension: 1'),
t('Existence: 9 Extension: 1'),
t('Existence: 8 Extension: 1'),
)),
t('PAM70') => drupal_map_assoc(array(
t('Existence: 8 Extension: 2'),
t('Existence: 7 Extension: 2'),
t('Existence: 6 Extension: 2'),
t('Existence: 11 Extension: 1'),
t('Existence: 10 Extension: 1'),
t('Existence: 9 Extension: 1'),
)),
t('PAM250') => drupal_map_assoc(array(
t('Existence: 15 Extension: 3'),
t('Existence: 14 Extension: 3'),
t('Existence: 13 Extension: 3'),
t('Existence: 12 Extension: 3'),
t('Existence: 11 Extension: 3'),
t('Existence: 17 Extension: 2'),
t('Existence: 16 Extension: 2'),
t('Existence: 15 Extension: 2'),
t('Existence: 14 Extension: 2'),
t('Existence: 13 Extension: 2'),
t('Existence: 21 Extension: 1'),
t('Existence: 20 Extension: 1'),
t('Existence: 19 Extension: 1'),
t('Existence: 18 Extension: 1'),
t('Existence: 17 Extension: 1'),
)),
t('BLOSUM80') => drupal_map_assoc(array(
t('Existence: 8 Extension: 2'),
t('Existence: 7 Extension: 2'),
t('Existence: 6 Extension: 2'),
t('Existence: 11 Extension: 1'),
t('Existence: 10 Extension: 1'),
t('Existence: 9 Extension: 1'),
)),
t('BLOSUM62') => drupal_map_assoc(array(
t('Existence: 11 Extension: 2'),
t('Existence: 10 Extension: 2'),
t('Existence: 9 Extension: 2'),
t('Existence: 8 Extension: 2'),
t('Existence: 7 Extension: 2'),
t('Existence: 6 Extension: 2'),
t('Existence: 13 Extension: 1'),
t('Existence: 12 Extension: 1'),
t('Existence: 11 Extension: 1'),
t('Existence: 10 Extension: 1'),
t('Existence: 9 Extension: 1'),
)),
t('BLOSUM45') => drupal_map_assoc(array(
t('Existence: 13 Extension: 3'),
t('Existence: 12 Extension: 3'),
t('Existence: 11 Extension: 3'),
t('Existence: 10 Extension: 3'),
t('Existence: 15 Extension: 2'),
t('Existence: 14 Extension: 2'),
t('Existence: 13 Extension: 2'),
t('Existence: 12 Extension: 2'),
t('Existence: 19 Extension: 1'),
t('Existence: 18 Extension: 1'),
t('Existence: 17 Extension: 1'),
t('Existence: 16 Extension: 1'),
)),
t('BLOSUM50') => drupal_map_assoc(array(
t('Existence: 13 Extension: 3'),
t('Existence: 12 Extension: 3'),
t('Existence: 11 Extension: 3'),
t('Existence: 10 Extension: 3'),
t('Existence: 9 Extension: 3'),
t('Existence: 16 Extension: 2'),
t('Existence: 15 Extension: 2'),
t('Existence: 14 Extension: 2'),
t('Existence: 13 Extension: 2'),
t('Existence: 12 Extension: 2'),
t('Existence: 19 Extension: 1'),
t('Existence: 18 Extension: 1'),
t('Existence: 17 Extension: 1'),
t('Existence: 16 Extension: 1'),
t('Existence: 15 Extension: 1'),
)),
t('BLOSUM90') => drupal_map_assoc(array(
t('Existence: 9 Extension: 2'),
t('Existence: 8 Extension: 2'),
t('Existence: 7 Extension: 2'),
t('Existence: 6 Extension: 2'),
t('Existence: 11 Extension: 1'),
t('Existence: 10 Extension: 1'),
t('Existence: 9 Extension: 1'),
)),
);
if (isset($options[$key])) {
return $options[$key];
} else {
return array();
}
}
/**
* Respond to Ajax dropdown call
*/
function ajax_example_dependent_dropdown_callback($form, $form_state) {
return $form['ALG']['SParam']['gapCost'];
}