# HG changeset patch # User devteam # Date 1400517300 14400 # Node ID b52b9c7aabd9aaebc0430dfd2256ad72a20f7d6d Imported from capsule None diff -r 000000000000 -r b52b9c7aabd9 short_reads_boxplot.png Binary file short_reads_boxplot.png has changed diff -r 000000000000 -r b52b9c7aabd9 short_reads_figure_score.py --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/short_reads_figure_score.py Mon May 19 12:35:00 2014 -0400 @@ -0,0 +1,248 @@ +#!/usr/bin/env python +""" +boxplot: +- box: first quartile and third quartile +- line inside the box: median +- outlier: 1.5 IQR higher than the third quartile or 1.5 IQR lower than the first quartile + IQR = third quartile - first quartile +- The smallest/largest value that is not an outlier is connected to the box by with a horizontal line. +""" + +import os, sys, math, tempfile, re +from rpy import * + +assert sys.version_info[:2] >= ( 2, 4 ) + +def stop_err( msg ): + sys.stderr.write( "%s\n" % msg ) + sys.exit() + +def merge_to_20_datapoints( score ): + number_of_points = 20 + read_length = len( score ) + step = int( math.floor( ( read_length - 1 ) * 1.0 / number_of_points ) ) + scores = [] + point = 1 + point_sum = 0 + step_average = 0 + score_points = 0 + + for i in xrange( 1, read_length ): + if i < ( point * step ): + point_sum += int( score[i] ) + step_average += 1 + else: + point_avg = point_sum * 1.0 / step_average + scores.append( point_avg ) + point += 1 + point_sum = 0 + step_average = 0 + if step_average > 0: + point_avg = point_sum * 1.0 / step_average + scores.append( point_avg ) + if len( scores ) > number_of_points: + last_avg = 0 + for j in xrange( number_of_points - 1, len( scores ) ): + last_avg += scores[j] + last_avg = last_avg / ( len(scores) - number_of_points + 1 ) + else: + last_avg = scores[-1] + score_points = [] + for k in range( number_of_points - 1 ): + score_points.append( scores[k] ) + score_points.append( last_avg ) + return score_points + +def __main__(): + + invalid_lines = 0 + + infile_score_name = sys.argv[1].strip() + outfile_R_name = sys.argv[2].strip() + + infile_name = infile_score_name + + # Determine tabular or fasta format within the first 100 lines + seq_method = None + data_type = None + for i, line in enumerate( file( infile_name ) ): + line = line.rstrip( '\r\n' ) + if not line or line.startswith( '#' ): + continue + if data_type == None: + if line.startswith( '>' ): + data_type = 'fasta' + continue + elif len( line.split( '\t' ) ) > 0: + fields = line.split() + for score in fields: + try: + int( score ) + data_type = 'tabular' + seq_method = 'solexa' + break + except: + break + elif data_type == 'fasta': + fields = line.split() + for score in fields: + try: + int( score ) + seq_method = '454' + break + except: + break + if i == 100: + break + + if data_type is None: + stop_err( 'This tool can only use fasta data or tabular data.' ) + if seq_method is None: + stop_err( 'Invalid data for fasta format.') + + # Determine fixed length or variable length within the first 100 lines + read_length = 0 + variable_length = False + if seq_method == 'solexa': + for i, line in enumerate( file( infile_name ) ): + line = line.rstrip( '\r\n' ) + if not line or line.startswith( '#' ): + continue + scores = line.split('\t') + if read_length == 0: + read_length = len( scores ) + if read_length != len( scores ): + variable_length = True + break + if i == 100: + break + elif seq_method == '454': + score = '' + for i, line in enumerate( file( infile_name ) ): + line = line.rstrip( '\r\n' ) + if not line or line.startswith( '#' ): + continue + if line.startswith( '>' ): + if len( score ) > 0: + score = score.split() + if read_length == 0: + read_length = len( score ) + if read_length != len( score ): + variable_length = True + break + score = '' + else: + score = score + ' ' + line + if i == 100: + break + + if variable_length: + number_of_points = 20 + else: + number_of_points = read_length + read_length_threshold = 100 # minimal read length for 454 file + score_points = [] + score_matrix = [] + invalid_scores = 0 + + if seq_method == 'solexa': + for i, line in enumerate( open( infile_name ) ): + line = line.rstrip( '\r\n' ) + if not line or line.startswith( '#' ): + continue + tmp_array = [] + scores = line.split( '\t' ) + for bases in scores: + nuc_errors = bases.split() + try: + nuc_errors[0] = int( nuc_errors[0] ) + nuc_errors[1] = int( nuc_errors[1] ) + nuc_errors[2] = int( nuc_errors[2] ) + nuc_errors[3] = int( nuc_errors[3] ) + big = max( nuc_errors ) + except: + #print 'Invalid numbers in the file. Skipped.' + invalid_scores += 1 + big = 0 + tmp_array.append( big ) + score_points.append( tmp_array ) + elif seq_method == '454': + # skip the last fasta sequence + score = '' + for i, line in enumerate( open( infile_name ) ): + line = line.rstrip( '\r\n' ) + if not line or line.startswith( '#' ): + continue + if line.startswith( '>' ): + if len( score ) > 0: + score = ['0'] + score.split() + read_length = len( score ) + tmp_array = [] + if not variable_length: + score.pop(0) + score_points.append( score ) + tmp_array = score + elif read_length > read_length_threshold: + score_points_tmp = merge_to_20_datapoints( score ) + score_points.append( score_points_tmp ) + tmp_array = score_points_tmp + score = '' + else: + score = "%s %s" % ( score, line ) + if len( score ) > 0: + score = ['0'] + score.split() + read_length = len( score ) + if not variable_length: + score.pop(0) + score_points.append( score ) + elif read_length > read_length_threshold: + score_points_tmp = merge_to_20_datapoints( score ) + score_points.append( score_points_tmp ) + tmp_array = score_points_tmp + + # reverse the matrix, for R + for i in range( number_of_points - 1 ): + tmp_array = [] + for j in range( len( score_points ) ): + try: + tmp_array.append( int( score_points[j][i] ) ) + except: + invalid_lines += 1 + score_matrix.append( tmp_array ) + + # generate pdf figures + #outfile_R_pdf = outfile_R_name + #r.pdf( outfile_R_pdf ) + outfile_R_png = outfile_R_name + r.bitmap( outfile_R_png ) + + title = "boxplot of quality scores" + empty_score_matrix_columns = 0 + for i, subset in enumerate( score_matrix ): + if not subset: + empty_score_matrix_columns += 1 + score_matrix[i] = [0] + + if not variable_length: + r.boxplot( score_matrix, xlab="location in read length", main=title ) + else: + r.boxplot( score_matrix, xlab="position within read (% of total length)", xaxt="n", main=title ) + x_old_range = [] + x_new_range = [] + step = read_length_threshold / number_of_points + for i in xrange( 0, read_length_threshold, step ): + x_old_range.append( ( i / step ) ) + x_new_range.append( i ) + r.axis( 1, x_old_range, x_new_range ) + r.dev_off() + + if invalid_scores > 0: + print 'Skipped %d invalid scores. ' % invalid_scores + if invalid_lines > 0: + print 'Skipped %d invalid lines. ' % invalid_lines + if empty_score_matrix_columns > 0: + print '%d missing scores in score_matrix. ' % empty_score_matrix_columns + + r.quit(save = "no") + +if __name__=="__main__":__main__() diff -r 000000000000 -r b52b9c7aabd9 short_reads_figure_score.xml --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/short_reads_figure_score.xml Mon May 19 12:35:00 2014 -0400 @@ -0,0 +1,87 @@ + + + + fontconfig + rpy + R + +short_reads_figure_score.py $input1 $output1 + + + + + + + + + + + + + + + + + + + + + +.. class:: warningmark + +To use this tool, your dataset needs to be in the *Quality Score* format. Click the pencil icon next to your dataset to set the datatype to *Quality Score* (see below for examples). + +----- + +**What it does** + +This tool takes Quality Files generated by Roche (454), Illumina (Solexa), or ABI SOLiD machines and builds a graph showing score distribution like the one below. Such graph allows you to perform initial evaluation of data quality in a single pass. + +----- + +**Examples of Quality Data** + +Roche (454) or ABI SOLiD data:: + + >seq1 + 23 33 34 25 28 28 28 32 23 34 27 4 28 28 31 21 28 + +Illumina (Solexa) data:: + + -40 -40 40 -40 -40 -40 -40 40 + +----- + +**Output example** + +Quality scores are summarized as boxplot (Roche 454 FLX data): + +.. image:: short_reads_boxplot.png + +where the **X-axis** is coordinate along the read and the **Y-axis** is quality score adjusted to comply with the Phred score metric. Units on the X-axis depend on whether your data comes from Roche (454) or Illumina (Solexa) and ABI SOLiD machines: + + - For Roche (454) X-axis (shown above) indicates **relative** position (in %) within reads as this technology produces reads of different lengths; + - For Illumina (Solexa) and ABI SOLiD X-axis shows **absolute** position in nucleotides within reads. + +Every box on the plot shows the following values:: + + o <---- Outliers + o + -+- <---- Upper Extreme Value that is no more + | than box length away from the box + | + +--+--+ <---- Upper Quartile + | | + +-----+ <---- Median + | | + +--+--+ <---- Lower Quartile + | + | + -+- <---- Lower Extreme Value that is no more + than box length away from the box + o <---- Outlier + + + + + diff -r 000000000000 -r b52b9c7aabd9 test-data/454.qual --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/test-data/454.qual Mon May 19 12:35:00 2014 -0400 @@ -0,0 +1,52 @@ +>EYKX4VC01B65GS length=54 xy=0784_1754 region=1 run=R_2007_11_07_16_15_57_ +33 23 34 25 28 28 28 32 23 34 27 4 28 28 31 21 28 27 27 28 28 28 28 28 28 28 33 23 28 33 24 36 27 31 21 28 28 33 26 33 24 27 28 28 28 28 28 28 28 32 23 28 34 25 +>EYKX4VC01BNCSP length=187 xy=0558_3831 region=1 run=R_2007_11_07_16_15_57_ +27 35 26 25 37 28 37 28 25 28 27 36 27 28 36 27 28 28 27 36 27 30 19 27 28 36 28 23 36 27 27 28 27 27 28 37 29 27 26 27 24 24 36 27 26 28 36 28 24 25 21 28 24 28 26 34 25 26 43 36 +22 9 23 19 28 28 28 28 26 28 39 32 12 27 36 28 28 26 37 28 28 26 28 28 28 27 28 26 36 27 27 27 36 28 27 27 28 28 36 27 36 28 39 32 12 35 28 26 37 29 28 28 28 28 37 29 28 36 28 35 +26 27 37 29 28 26 28 36 28 26 24 38 32 11 28 26 32 24 36 32 18 2 27 25 33 26 32 28 6 18 22 26 17 15 14 28 20 8 22 21 14 22 26 16 26 16 28 20 22 27 18 27 18 27 28 27 20 25 34 27 +27 33 25 34 28 8 26 +>EYKX4VC01CD9FT length=115 xy=0865_1719 region=1 run=R_2007_11_07_16_15_57_ +35 24 16 9 2 27 39 33 13 36 27 36 27 28 28 28 27 28 28 33 23 37 28 28 28 36 27 28 28 28 28 36 27 28 28 28 27 28 28 28 28 28 28 28 37 28 28 28 28 37 28 26 28 36 27 28 28 28 28 28 +28 28 28 37 28 28 35 26 27 28 28 27 36 27 35 25 32 22 28 28 28 28 28 28 28 28 28 34 25 36 27 34 25 28 27 28 28 36 27 28 35 29 6 28 28 28 37 30 8 33 24 28 27 27 27 +>EYKX4VC01B8FW0 length=95 xy=0799_0514 region=1 run=R_2007_11_07_16_15_57_ +28 40 34 15 35 28 6 27 33 23 34 24 32 22 28 28 28 41 34 17 28 28 28 37 30 9 28 28 28 28 28 27 28 37 30 8 28 28 27 28 35 26 27 35 26 28 37 30 9 28 27 28 28 28 34 25 28 28 32 22 +26 28 28 28 28 27 43 36 23 12 1 28 21 28 27 16 28 32 23 27 28 27 28 28 27 28 28 28 32 22 28 26 26 27 28 +>EYKX4VC01BCGYW length=115 xy=0434_3926 region=1 run=R_2007_11_07_16_15_57_ +28 6 26 15 27 28 37 28 41 35 17 28 21 28 23 21 27 36 27 24 36 28 40 34 14 22 25 28 24 27 28 37 28 26 28 27 27 28 28 28 28 27 43 36 22 8 28 26 28 27 26 14 28 25 20 28 34 24 25 40 +33 18 1 19 27 16 36 28 36 28 21 27 25 41 34 16 22 28 37 29 26 26 35 27 28 26 41 34 16 28 28 27 28 37 29 25 43 36 23 12 1 11 39 32 12 28 17 20 28 28 17 36 29 7 24 +>EYKX4VC01AZXC6 length=116 xy=0292_0280 region=1 run=R_2007_11_07_16_15_57_ +35 24 17 11 5 26 24 40 33 14 34 25 33 24 28 27 27 26 28 28 33 24 36 27 28 27 36 27 27 28 27 28 35 26 27 27 28 27 28 28 28 28 28 27 28 36 27 28 28 28 37 28 27 26 35 26 27 28 28 28 +27 28 28 28 37 29 28 35 26 28 27 28 28 35 26 35 26 31 21 28 28 28 28 28 28 28 28 28 34 25 35 26 35 26 28 28 28 28 37 28 27 37 30 9 28 28 28 37 30 9 33 23 28 28 28 27 +>EYKX4VC01CATH5 length=82 xy=0826_0843 region=1 run=R_2007_11_07_16_15_57_ +28 28 41 35 17 33 24 28 27 28 28 28 36 27 28 28 28 36 27 34 25 27 28 28 28 28 28 28 28 37 28 27 35 25 28 26 28 27 28 28 24 36 27 26 26 37 30 9 28 28 28 28 28 25 25 35 26 26 27 35 +25 28 36 28 28 28 31 21 25 13 32 22 41 34 17 0 22 10 32 23 24 28 +>EYKX4VC01BCEIV length=47 xy=0434_0757 region=1 run=R_2007_11_07_16_15_57_ +28 28 26 26 28 26 28 27 28 25 32 22 27 26 25 27 28 28 27 26 27 28 32 23 28 28 34 25 27 22 26 26 27 28 17 28 28 28 28 28 28 34 24 35 25 28 28 +>EYKX4VC01BWERM length=83 xy=0662_0304 region=1 run=R_2007_11_07_16_15_57_ +28 36 27 28 28 35 26 34 24 28 28 28 32 22 28 28 32 23 28 28 32 23 27 28 27 34 27 3 27 43 36 22 9 35 26 37 29 26 27 32 23 28 28 27 28 36 27 28 36 27 28 28 28 28 28 35 26 34 25 28 +36 30 8 28 28 28 28 27 27 28 28 28 28 37 28 28 36 27 28 39 33 13 27 +>EYKX4VC01BT2O7 length=69 xy=0635_1945 region=1 run=R_2007_11_07_16_15_57_ +28 28 28 28 41 34 17 27 28 31 21 28 27 32 23 36 27 28 28 33 24 28 27 28 28 27 32 22 28 34 27 3 27 43 36 22 8 27 28 34 27 3 28 28 28 28 28 28 28 33 23 28 28 28 28 34 24 28 34 24 +28 28 27 36 27 28 37 30 9 +>EYKX4VC01BO0UO length=222 xy=0577_3838 region=1 run=R_2007_11_07_16_15_57_ +27 27 28 36 27 28 39 33 13 28 28 28 27 28 37 28 28 41 35 17 28 28 28 27 28 26 36 27 28 36 27 27 28 27 35 26 27 26 28 28 28 28 28 36 27 28 28 38 31 10 24 27 27 27 27 27 28 28 37 28 +27 28 35 26 28 28 36 27 28 28 27 28 28 28 28 28 28 27 36 28 27 36 27 37 28 27 28 27 28 28 28 27 28 28 27 36 27 26 27 28 28 28 28 28 37 28 37 29 25 28 36 27 28 27 28 34 27 26 24 34 +28 28 28 31 23 27 28 34 27 28 37 33 14 23 37 33 15 38 34 23 13 2 26 24 28 26 35 31 12 36 32 14 31 22 24 28 27 33 26 26 27 27 27 27 28 27 35 30 11 26 27 35 31 12 28 27 26 27 36 32 +14 27 34 27 37 33 15 27 27 34 28 27 23 27 35 31 11 27 28 28 26 34 26 27 28 34 28 28 28 39 35 22 9 27 27 23 27 35 28 34 27 27 +>EYKX4VC01CBCPK length=83 xy=0832_1158 region=1 run=R_2007_11_07_16_15_57_ +28 35 26 28 28 35 26 35 26 28 28 28 34 24 28 28 35 25 28 28 34 25 28 28 27 35 28 5 28 43 36 22 9 35 26 37 28 28 27 32 23 27 28 28 28 36 27 28 36 27 28 28 28 28 28 36 27 35 25 28 +37 30 9 28 28 28 28 28 28 28 28 28 28 36 27 28 35 26 28 38 31 10 28 +>EYKX4VC01B474S length=54 xy=0762_2010 region=1 run=R_2007_11_07_16_15_57_ +28 28 28 28 27 43 36 23 11 33 23 27 25 26 28 28 39 33 13 28 27 29 18 28 26 27 26 28 27 28 36 27 26 28 28 28 28 25 28 41 34 17 24 36 28 37 28 28 26 28 17 27 28 26 +>EYKX4VC01BB4QL length=57 xy=0431_0363 region=1 run=R_2007_11_07_16_15_57_ +36 24 15 7 28 33 26 28 27 27 26 29 18 28 35 26 28 26 28 25 27 27 28 36 27 41 34 20 5 27 36 28 28 28 27 28 32 22 34 25 28 28 28 28 26 28 27 28 36 27 40 34 18 3 28 37 28 +>EYKX4VC01BJ37M length=64 xy=0522_0192 region=1 run=R_2007_11_07_16_15_57_ +28 26 28 28 28 28 28 28 27 28 28 27 27 36 27 37 28 28 28 28 28 28 28 28 28 27 36 29 8 39 33 13 28 36 27 41 34 20 5 28 28 28 27 36 28 28 28 28 27 28 28 28 28 27 37 30 8 27 28 26 +33 26 35 26 +>EYKX4VC01BV9R8 length=54 xy=0660_2038 region=1 run=R_2007_11_07_16_15_57_ +41 34 19 4 27 28 28 30 20 28 28 34 27 4 28 28 27 34 25 27 28 28 28 28 28 28 28 28 28 28 28 28 38 31 11 27 28 28 28 28 37 28 40 33 18 2 24 15 25 24 12 26 34 27 +>EYKX4VC01CEPP8 length=60 xy=0870_2350 region=1 run=R_2007_11_07_16_15_57_ +26 21 40 34 17 26 36 29 8 26 28 22 26 28 28 20 24 28 34 26 23 11 28 28 26 27 26 40 33 14 27 35 26 26 23 10 28 31 21 28 23 27 23 28 36 27 26 36 28 27 36 28 28 27 25 27 27 27 26 28 +>EYKX4VC01BTLME length=78 xy=0630_0292 region=1 run=R_2007_11_07_16_15_57_ +36 27 25 24 33 23 28 26 28 28 28 28 27 27 26 36 27 36 28 28 28 36 27 28 28 27 27 36 28 28 37 29 28 26 36 27 27 28 27 27 28 27 36 28 28 28 36 27 36 27 28 36 28 25 36 28 28 28 27 28 +39 33 13 28 28 37 28 28 41 34 16 28 28 28 26 36 28 24 diff -r 000000000000 -r b52b9c7aabd9 test-data/454Score.png Binary file test-data/454Score.png has changed diff -r 000000000000 -r b52b9c7aabd9 test-data/solexa.qual --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/test-data/solexa.qual Mon May 19 12:35:00 2014 -0400 @@ -0,0 +1,30 @@ + -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 -15 15 -40 40 -40 -40 + -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -5 5 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 21 -40 -21 -40 40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 12 -40 -40 -12 -36 -40 36 -40 -40 -40 40 -40 -4 4 -40 -40 -40 -40 -40 40 -40 -40 14 -14 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -25 25 + -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 34 -40 -34 40 -40 -40 -40 -40 -40 -40 40 -40 -25 25 -40 -40 -40 -40 40 -37 -40 37 -40 -40 7 -40 -7 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 40 -40 -40 -40 38 -40 -40 -38 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 11 -16 -13 -22 -40 -40 40 -40 -40 -40 -40 40 + 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -33 33 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -25 25 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 27 -27 -5 5 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 -37 37 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -25 25 40 -40 -40 -40 -40 -40 34 -34 -40 40 -40 -40 + -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -2 2 -40 -40 35 -35 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 -40 36 -40 -36 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 5 -5 -40 -28 -40 -16 -40 16 -40 40 -40 -40 + -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 28 -28 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 + -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 27 -27 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 14 -14 -40 40 -40 -40 + -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -36 -40 -40 36 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 + -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 22 -22 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 + -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -6 6 -40 40 -40 -40 -40 -40 -40 40 3 -40 -40 -3 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -0 -40 -40 0 -40 -40 -40 40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -20 20 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 8 -40 -40 -8 -40 40 -40 -40 -40 -40 40 -40 -4 4 -40 -40 + -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -1 1 -40 -40 40 -40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 28 -28 -40 -40 -40 40 -10 -40 10 -40 -40 40 -40 -40 40 -40 -40 -40 -40 2 -2 -40 40 -40 -40 -40 17 -40 -17 -40 -40 40 -40 -40 3 -3 -40 -40 -40 -40 -40 40 -28 -40 28 -40 -40 -40 40 -40 13 -26 -13 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 18 -18 40 -40 -40 -40 -10 -40 10 -40 -40 -40 40 -40 17 -40 -17 -40 -28 28 -40 -40 + -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 5 -40 -40 -5 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -10 10 -40 -40 -40 -40 40 -40 -40 28 -28 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 32 -32 -40 -40 -40 4 -40 -4 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 + -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 27 -27 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 + -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -15 15 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 32 -40 -40 -32 -40 -40 -40 40 + -40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 + -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 25 -40 -40 -25 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -20 20 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -31 31 -40 -40 -40 -40 -40 40 -40 -23 23 -40 -40 -40 -40 40 -40 -40 40 -40 15 -15 -40 -40 -40 40 -40 -40 25 -40 -25 -40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 39 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 + -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 + -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 + -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 -40 40 + -40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 + -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -36 -40 36 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -24 -40 -40 24 -40 -40 -40 40 -14 -40 -40 14 -40 -40 -40 40 -40 -40 -40 40 + -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 9 -9 -40 -40 -40 40 23 -40 -40 -23 + -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 20 -20 40 -40 -40 -40 -40 -10 10 -40 14 -40 -40 -14 -40 -40 40 -40 -40 -40 -40 40 -35 -40 35 -40 -40 38 -38 -40 40 -40 -40 -40 -40 -40 13 -13 -40 -40 40 -40 -40 -40 -40 40 29 -29 -40 -40 -40 -40 -40 40 -15 -40 -40 15 40 -40 -40 -40 40 -40 -40 -40 -23 -40 -40 23 40 -40 -40 -40 -1 -40 -40 1 -40 38 -38 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -9 9 -40 -40 -21 21 -40 -40 21 -21 -40 -40 -29 29 -40 -13 13 -40 40 -40 -40 -40 -40 -40 40 -40 + -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -3 3 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 12 -40 -40 -12 -10 -40 10 -40 -40 -40 40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 -21 -40 -40 21 + -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 + -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -1 1 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -31 31 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 + -40 -40 40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 + -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 40 -40 -40 -29 29 -40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 36 -36 40 -40 -40 -40 + -40 -40 40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 20 -20 -40 -40 -40 -40 -40 40 3 -40 -40 -3 40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 -40 -40 33 -33 40 -40 -40 -40 -13 -40 -40 13 8 -8 -40 -40 -40 28 -28 -40 -29 -40 -40 29 -40 -40 -40 40 2 -40 -2 -26 -40 -28 27 -33 -1 0 -40 -35 31 -31 -40 -40 -40 40 -40 -40 -11 -40 -40 11 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 + -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 40 40 -40 -40 -40 -40 40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 38 -38 -40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 40 -40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 -40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 40 -40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 40 -40 -40 -40 -40 -40 -40 40 -40 -40 -40 40 29 -40 -40 -29 -40 40 -40 -40 -40 10 -40 -10 40 -40 -40 -40 -40 -40 2 -2 -13 -40 -40 13 -40 -40 40 -40 -40 -40 -40 40 -40 -40 -40 40 diff -r 000000000000 -r b52b9c7aabd9 test-data/solexaScore.png Binary file test-data/solexaScore.png has changed diff -r 000000000000 -r b52b9c7aabd9 tool_dependencies.xml --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/tool_dependencies.xml Mon May 19 12:35:00 2014 -0400 @@ -0,0 +1,12 @@ + + + + + + + + + + + +