Rcount: User Guide - Institute of Plant Biology
Contents
1. bit binaries for Linux Windows and Mac and R Scripts can be downloaded on www botinst uzh ch research development grossnik rcount html The pro grams can easily use more than 3 Gb of RAM and it is therefore strongly recommended to use a 64 bit system with at least 6 Gb of RAM see Table 4 for examples concerning the memory usage 1 1 Using pre compiled binaries If you have a 64 bit Ubuntu like Linux Windows 7 or MacOSX use the pre compiled binaries The binaries were built on Kubuntu 12 04 Windows 7 and MacOS 10 9 versions below 10 9 were not tested If you encounter problems with the binaries try building the programs from source see section 5 and send a report to marcschmid gmx ch 1 1 1 Linux Download and unpack the archive linux 64bit zip Start Rcount multireads Rcount format and Rcount distribute directly either by double clicking on them or from the terminal you may need to make them executable first right click on the binaries open the properties dialog and check the box for is executable or in a terminal type chmod 755 filename 1 1 2 Windows Download and unpack the archive windows_64bit zip Start the applications directly by double clicking on them 1 1 3 Mac Download and unpack the archive mac_64bit zip Mount the dmg files double click and start the applications by double clicking on them 2 Step by step example This section provides a step by step tutorial on how to get count2. The panel on the top right compares the gene expression values of the two samples i e one dot reflects a gene Colors indicate the point density red and blue indicate the highest respectively lowest densities The panel on the lower left compares the sorted expression values of the two samples i e one dot reflects the x th highest expression value within one sample The latter gives an indication of how sim ilar the two distributions of expression values are e g if a certain value means the same in both samples Expression values correspond to log2 counts 1 CorP Pearson correlation CorS Spearman correlation n number of genes plotted 3 Usage This section describes the features requirements and parameters of Rcount in more detail 3 1 Aligning the short reads After initial quality checks have been performed e g with the FastQC soft ware the reads are aligned to a reference genome preferentially with a splice aware aligner In the step by step example we used TopHat2 3 to align the reads In principle any aligner works with Rcount but some may require the bam file to be additionally processed before using it The following features of the bam file are crucial for Rcount e The file must be sorted and an index must be present Rcount multireads needs both sorting and indexing Rcount distribute does not need an index but sorting is strongly recommended Rcount multireads keeps the order intact resorting is
3. cus B Locus A shall be truly expressed i e is in vivo expressed and therefore has reads mapping to it locus B not i e all reads that map to B are coming from A Using a single step all hits of locus A would be declared as ambiguous In case locus B has no unambiguous hit the hits from locus A would be equally distributed to locus A and B leading to an underestimation of the expression value from locus A and an overestimation of the expression value from locus B a false positive In another case where locus B has one or two unambiguous hits due to sequencing and or alignment errors all the hits from locus A would be wrongly assigned to locus B one false positive and one false negative The same error would occur if locus A has a longer transcript in vivo than the in silico genome annotation would indicate The hits at the borders of locus A would then be unambiguously assigned to locus B and as a consequence also all the ambiguous hits To overcome this scenario at least to some extent Rcount distribute uses a two step approach In the first step all hits are mapped to all annotated transcripts In many cases one can expect that each truly ex pressed transcript should have at least some hits within the first few bps at its 3 end because the library preparation protocols frequently rely on poly A tail priming for cDNA synthesis In addition one can set a minimal number of hits Transcripts not matching these criteria are the
4. families with similar sequence To avoid underestimation of expression values of transcripts with similar sequence one can allow multiple alignments of one read to several locations in the genome However a read r with m gt 1 alignments would count m times resulting in overestimation of expression values To overcome this Rcount multireads calculates for each alignment of such a read the weight H using a score S divided by the sum of scores from all alignments of the read H S 5757 Si If So Si is zero all alignments of the respective read are discarded i e all weights set to zero For ungapped alignments the score is defined as sum of coverage originating from uniquely aligned reads at the position of the alignment and the surrounding region The size of the region can be set by choosing the allocation distance This value will be added on both sides of the alignment position For gapped alignments the score equals to the number of uniquely aligned reads spanning the same gap Thus if a read has both types of align ments the ungapped ones are generally preferred Further notes e Rcount multireads is not required for Rcount distribute to run If you do not use reads with multiple alignments you can safely skip Rcount multireads It is important that the number of unique alignments is well above the number of multireads If not it is better to use only the uniquely aligned reads high number of multireads was
5. observed especially in amplified libraries where the cDNA was prepared using random primers in house data e Memory usage is mainly influenced by the number of multireads The main memory consumption of Rcount multireads is caused by storing the multireads The size of the reference genome does not make a big difference If RAM is limiting one could lower the maximum number of alignments per read e Known issues on Mac and Windows The progress bar is not updated instantly It instead advances directly to 80 after finishing reading the files 10 3 4 Rcount distribute Mapping aligned reads to genomic features To count the number of reads per genomic feature the positions of the aligned reads termed hits have to be mapped to the genomic features Hits are matched to the genomic features as follows ungapped hits must be entirely within one exon of a given gene and gapped hits must accurately span exon junctions corresponds to the intersection_strict mode in HTSeq Hits can be divided into unambiguous mapping to transcripts of only one locus and ambiguous mapping to transcripts of more than one locus To avoid count ing ambiguous hits multiple times Rcount distribute proportionally distributes them based on the number of unambiguous hits If there are no unambiguous hits the ambiguous hits are equally distributed However we assume a case where two loci A and B overlap such that locus A is entirely located within lo
6. the downstream analysis programs counts lt f read Rcount pathToWorkingDirectory save a merged table outfile lt file path pathToWorkingDirectory all_counts csv write csv counts outfile draw pairwise scatter plots unsorted sorted and histograms f pair all log2 countst1 pathToWorkingDirectory Table 1 The count table produced in the step by step example type head counts in R to see this SRR976339_counts SRR976340_counts LOC_Os01g01019 13 10 LOC_Os01g01030 55 35 LOC_Os01g01040 127 111 LOC_Os01g01060 70 47 LOC_Os01g01080 5 10 LOC_Os01g01115 28 10 20 p S median 5 585 corP 0 824 F 0 15 mean 5 897 corS 0 893 i sd 2 322 n 14867 15 20 2 0 10 5 o 9 8 10 eo r Fa f re E oO 0 05 5 000 Se o J Sa 0 5 10 15 20 0 5 10 15 SRR976339_counts SRR976340_counts 20 corp 0 971 A median 5 392 corS 0 997 mean 5 746 n 14867 0 15 sd 2 266 15 o f 5 2 8 0 10 l 10 8 E Q T T 2 0 05 5 0 7 0 00 A 0 5 10 15 20 0 5 10 15 SRR976339_counts SRR976340_counts Figure 1 Distribution and pair wise comparison of gene expression of the two samples processed in the step by step tutorial The plots were generated with the function f pair all included in the R script Top left and bottom right pan els show the distributions of expression values for each sample
7. Recount User Guide Marc W Schmid marcschmid gmx ch June 25 2014 Contents 1 Installation 1 1 1 Using pre compiled binaries 1 Le mse oo 4 a Reo ccd wm BS ease eee ce ete de be 1 14 22 Windows se Sers pisa oe lke REE oe Bie me vas 1 Ade MaG uont eS x Ae EAS ahah CA DE ee HP Se Rea Bin A A el og AT A 1 Step by step example 2 2 1 OPTIONAL Installation of additional programs 2 2 2 OPTIONAL Obtaining the short read data 3 2 3 OPTIONAL Aligning the short reads to the reference genome 3 24 Weighting reads that have more than one alignment with Rcount MUULEVT EGS oe 0 ee ES LE AM ae Rte rss Bae Sak 4 2 5 Reformating the genome annotation with Rcount format 4 2 6 Counting the number of hits per gene with Rcount distribute 4 2 7 Merging the count tablesinR 5 Usage 8 3 1 Aligning the short reads 8 3 2 Rcount format Creating the genome annotation file 9 3 3 Rcount multireads Weighting reads with more than one alignment 10 3 4 Rcount distribute Mapping aligned reads to genomic features 11 3 5 Reading and merging tables in R 12 Test data 13 Building from source 15 Dale AUS AAA Rt ae bh ME NE COS RE a ME Sal a 15 022 4 Windows cgi eee toe eee dE de Ro NE A A 15 Did MaS LA ck ee ani Ae ed ted ot Le RE Det ho the 16 1 Installation This section describes how to obtain and install Rcount Source code 64
8. atel Q Peng S Takeo R Kawahara Miki H Goto F Cao K Kimura Y Monji T Kuwayama H Iwata Age associated changes in gene expression and de velopmental competence of bovine oocytes and a possible countermeasure against age associated events Molecular Reproduction and Development 80 7 2013 508 521 M Washburn B Kakaradov B Sundararaman E Wheeler S Hoon G Yeo H H A The dsRBP and inactive editor ADR 1 utilizes dsRNA binding to regulate A to I RNA editing across the C elegans transcriptome Cell Reports 6 4 2014 599 607 G Ramaswami R Zhang R Piskol L Keegan P Deng M O Connell J Li Identifying RNA editing sites using RNA sequencing data alone Nature Methods 10 2 2013 128 132 Y Ku N Renaud R Veile C Helms C Voelker M Warchol M Lovett The transcriptome of utricle hair cell regeneration in the avian inner ear The Journal of Neuroscience 34 10 2014 3523 3535 J Odawara A Harada T Yoshimi K Maehara T Tachibana S Okada K Akashi Y Ohkawa The classification of mRNA expression levels by the phosphorylation state of RNAPII CTD based on a combined genome wide approach BMC Genomics 12 516 M Yukawa T Akiyama V Franke N Mise T Isagawa Y Suzuki M Suzuki K Vlahovicek K Abe H Aburatani F Aoki Genome Wide Analysis of the Chromatin Composition of Histone H2A and H3 Variants in Mouse Embryonic Stem Cells PLOS ONE 9 3 2014 e92689 D Cor
9. he user to edit some aspects of the genome annotation e Loci and their transcripts can be extended on both sides Some samples have an abnormal number of hits in non genic regions often caused by hits directly flanking the known loci Cases like this can be easily detected by comparing the results with and without the extension of the loci e Certain feature types can be removed by setting their priority to zero The genome annotation sometimes contains ambiguous information For example it often contains coding sequence and UTR information However these are an in silico specification of the type exon i e exon already includes UTR and CDS which behave identically during library preparation Thus they can be savely removed e Features can be given different priorities during the distribution of am biguous reads Depending on the library preparation protocol it is possible that some of the features in the genome annotation are very unlikely to be sequenced e g rRNA coding genes with poly A selective library preparation pro tocols Instead of removing these features entirely it is possible to set a lower priority to them In case a read aligns to a location where two features with different priorities overlap it is automatically assigned to the one with a higher priority It is important to note that priorities are only considered on a given level In case of a gene and a pseudogene top level in the genome annota
10. ial SRR976339 The count table should be stored in the working directory path to rice_tutorial Go to the Parameters tab and enable Use multiple alignments Leave the rest unchanged and click OK A dialog will open and ask you to save the project save it as SRR976339 xml in path to rice tutorial Once you have cre ated for each sample a project start the analysis by clicking run all Once a sample is processed the alignment statistics are automatically added to its project saving is done automatically as well 2 7 Merging the count tables in R Rcount distribute creates for each sample a separate count table stored in the working directory e g SRR976339_counts txt The provided R Script offers a function to read in all count tables from a given directory and merge them into a single table Open R and use the following code to obtain a merged table and to generate a plot that gives a first impression of the data correlation and expression value distribution table 1 and figure 1 set the path of the R script provided see in linux_64bit zip or windows_64bit zip or mac_64bit zip pathToScript lt path to Rcount R functions R set the path to the folder where the count tables are located pathToWorkingDirectory lt path to rice_tutorial load the functions from the script source pathToScript read in the count tables note that the expression values are rounded to integer numbers required by
11. ies gt General change the vari able Character Set to Use Multi Byte Character Set x under Configuration Properties gt C C gt Preprocessor in the field Preprocessor Definitions remove the UNICODE key words and add the key word _CRT_SECURE_NO_WARNINGS x under Configuration Properties gt C C gt Command Line add the options W2 wd4996 in the field with the Additional Options close the project properties and compile it with Build gt Build Solution 5 3 Mac Currently there are no build instructions available for the MacOS 16 References 1 S Anderson C Johnson D Jones L Conrad X Gou S Russell V Sun 10 11 12 daresan Transcriptomes of isolated Oryza sativa gametes characterized by deep sequencing evidence for distinct sex dependent chromatin and epige netic states before fertilization The Plant Journal 76 2013 729 741 B Langmead S Salzberg Fast gapped read alignment with Bowtie 2 Na ture Methods 9 2012 357 359 D Kim G Pertea C Trapnell H Pimentel R Kelley S Salzberg TopHat2 accurate alignment of transcriptomes in the presence of inser tions deletions and gene fusions Genome Biology 14 2013 R36 H Li B Handsaker A Wysoker F T J Ruan N Homer G Marth G Abecasis R Durbin G P D P Subgroup The Sequence align ment map SAM format and SAMtools Bioinformatics 25 2009 2078 2079 A Loraine S McCormick A Estrada K P
12. index SRR976339 fastq tophat p 6 g 10 no coverage search o SRR976340 rice_genome_index SRR976340 fastq The Rcount multireads program requires the bam files to have an index bai file Build this index with takes few time cd path to rice_tutorial SRR976339 samtools index accepted_hits bam cd path to rice_tutorial SRR976340 samtools index accepted_hits bam 2 4 Weighting reads that have more than one alignment with Rcount multireads With the TopHat2 option g 10 enabled it is possible that a read aligns to up to 10 locations in the genome To avoid counting the read 10 times we can weight these individual hits with Rcount multireads Start the program by clicking on it specify the input file accepted_hits bam and a corresponding output file e g accepted hits weighted bam Leave the allocation distance at 100 bp approximately one read length Press OK to start the weighting This will take some minutes Once the program has finished you can either process another file or close the program Run the program on both samples SRR976339 and SRR976340 2 5 Reformating the genome annotation with Rcount format The genome annotation has to be in a specific xml format You can create this file using Rcount format Start the program by clicking on it Specify the input genome annotation file all gff stored in your working directory leave the rest unchanged and press Next this can take up to a minute or longer on older system
13. itecture unpack it open a terminal and type adjust the path and version number SOLUTION 1 cd path to sratoolkit x x x x ubuntu64 bin sudo cp r usr local bin SOLUTION 2 temporary export PATH PATH path to sratoolkit x x x x ubuntu64 bin e Bowtie2 2 Visit bowtie bio sourceforge net bowtie2 index shtml and obtain the latest version Follow the link in the box on the right side of the page download the archive for linux unpack it open a terminal and type adjust the path and version number SOLUTION 1 cd path to bowtie2 x x x sudo cp bowtie2 usr local bin SOLUTION 2 temporary export PATH PATH path to bowtie2 x x x e TopHat2 3 Visit ccb jhu edu software tophat index shtml follow the link in the box on the right side of the page download the latest version for linux Linux x86_64 binary unpack it open a terminal and type adjust the path and version number SOLUTION 1 cd path to tophat 2 x x Linux_x86_64 rm README rm AUTHORS rm COPYING sudo cp usr local bin SOLUTION 2 temporary export PATH PATH path to tophat 2 x x Linux_x86_64 e SAMtools 4 Visit sourceforge net projects samtools files samtools and obtain the latest version Download the archive and unpack it SAM tools needs to be built from source For this install zlib zliblg zlibig dev and zlib1g dev from the package manager open a terminal and type adjust the path and ve
14. memory usage were obtained on a 64 bit Kubuntu 12 04 running on an Intel Core i7 930 2 8 GHz with 24 Gb RAM Input and output files were read and written from the same hard drive Samsung HD 7 200 rpm Memory usage of Rcount multireads is mostly in fluenced by the number of reads with multiple alignments e g high memory requirements for the mouse sample compared to the human sample However for Rcount distribute the size of the genome annotation has the largest impact on the memory usage e g the xml file holding the genome annotation of the human genome is with 643 Mb almost double the size compared to the xml genome annotation file for the mouse genome table 4 Table 3 Data used to test Recount organism assembly accession reads A thaliana TAIR10 SRX275909 5 52181949 B taurus Btau4 0 DRR001892 6 69040753 C elegans WS220 SRR1015366 7 34542067 C familiaris BROADD2 DRR001151 113253542 D melanogaster BDGP5 25 SRR629969 8 31267571 G gallus WASHUC2 SRR1264638 9 20 189 799 H sapiens GRCh37 DRRO000897 10 22635328 M musculus NCBIM37 DRR013118 11 27028925 R norvegicus RGSC3 4 SRR1041766 12 28 897 182 13 Table 4 Alignment statistics approximate run time and memory usage of Recount using the test data given in Table 3 Note that the number of mul tireads only includes the ones with at least one alignment with a weight above Zero organism reads aligned Rcount multireads Rcount dis
15. n discarded During the second step the hits are then divided into unambiguous and ambiguous The unambiguous hits are assigned first and used to distribute the ambiguous hits The transcripts are then filtered again using the same criteria as before The final expression value of a locus is then calculated as the sum of hits assigned to any of its transcripts Rcount distribute takes a genome file xml and an alignment file bam as input As a result it writes a new bam file with the mapping tag XM i x and a count table All parameters are set while creating a new project and stored in the project file Additionally one can enable the use of strand information if the library preparation protocol was strand specific note that Rcount checks for strand equality This greatly helps to assign the ambiguous reads as overlapping genes are often in opposite orientation to each other Note that in case of paired end reads only the first forward read is used 11 3 5 Reading and merging tables in R Rcount distribute creates for each sample a separate count table with multiple columns e sumUnAmb number of unambiguously mapped hits e sumAmb number of ambiguously mapped hits before distributing them accoring to the number of unambiguously mapped hits e sumAllo number of ambiguously mapped hits after distributing them accoring to the number of unambiguously mapped hits e sumDistUnAmb sumDistAmb sumDistAllo are identical to sumU
16. nAmb sumAmb sumAllo but only within the first x bp from the 3 of the tran script The value x can be specified in Rcount distribute by enabling consider 3 bias e TH total number of hits equals to sumUnAmb sumAllo The provided R Script offers a function to read in all table files from a given directory and merge them into a single table see section 2 7 The func tion f read Rcount loads per default the column TH Other columns in the count tables can be loaded using the argument toReturn of the function f read Rcount e g f read Rcount pathToWorkingDirectory toReturn sumUnAmb 12 4 Test data Recount was tested with data from multiple organisms with different genome sizes and RNA Seq libraries table 3 Genome annotation data and genome indices were obtained as described before see section 3 2 from tophat cbcb umd edu igenomes shtml Short reads were downloaded from www ncbi nlm nih gov sra The reformatted xml genome annotation as well as the resulting project files and count tables can by found on www botinst uzh ch research development grossnik rcount html in the archives test_data_annotations zip and test_data_results zip Reads were aligned with TopHat2 version 2 0 11 3 allowing up to 10 alignments per read g 10 For increasing the speed of TopHat2 the coverage search was turned off for all data Allocation distance for weighting multireads was set to 50 for all samples Run time and
17. paths e g sudo cp r seqan usr local include or add a line INCLUDEPATH lt path containing seqan folder gt into the pro files For Rcount format seqan is not required and this step can therefore be skipped finally in the QtCreater menu select Build gt Build Project my_project Windows Compiling on Windows requires access to VisualStudio The following steps worked well with VisualStudio 2010 Ultimate download and unpack the archive source zip install the Qt Add In for VisualStudio make sure to use Qt 4 x x libraries open each of the pro files Qt gt Open Qt Project File pro Rcount format source p502 format wizard p502 format wizard pro Rcount multireads source p502 process multireads p502 process multireads pro Rcount distribute source p502dataAnalysisRNA p502dataAnalysisRNA pro for Rcount format compile the code with Build gt Build Solution for Rcount multireads and Rcount distribute download the latest zlib from zlib net e g zlib 1 2 8 tar gz and unpack the archive open the project properties right click on the project in the solution explorer and do the following steps 15 x under Configuration Properties gt VC Directories in the field Include Directories add the zlib folder x under Configuration Properties gt VC Directories in the field Include Directories add the folder containing the seqan folder x under Configuration Propert
18. rsion number COMPULSORY BUILD INSTRUCTIONS cd path to samtools x x x make SOLUTION 1 cd path to samtools x x x sudo cp samtools usr local bin SOLUTION 2 temporary export PATH PATH path to samtools x x x 2 2 OPTIONAL Obtaining the short read data The data used in this tutorial can be conveniently retrieved from NCBI using the SRA toolkit Open a terminal to download the example data takes quite some time in the working directory e g rice_tutorial which has been automatically created by unpacking the archive rice_tutorial zip cd path to rice_tutorial fastq dump SRR976339 fastq dump SRR976340 2 3 OPTIONAL Aligning the short reads to the reference genome The alignment of the short reads to the reference genome using TopHat2 requires an index of the reference genome Build this index with takes quite some time cd path to rice_tutorial bowtie2 build f q o 0 all chrs con rice_genome_index You can now align the reads with TopHat2 Note that the option p 6 tells the computer to use six cores You may need to change this according to your sys tem The options g 10 and no coverage search are given to save memory and run time g 10 sets 10 as maximal number of alignments per read and no coverage search omits searching novel exon junctions using read cover age cd path to rice_tutorial mkdir SRR976339 mkdir SRR976340 tophat p 6 g 10 no coverage search o SRR976339 rice_genome_
19. s You can now inspect the structure of the genome annotation Click on gene to expand the menu If you expand the entry mRNA you see five sub entries CDS exon five_prime_UTR splice and three prime UTR Note that only exon and splice are important considering that the other three types CDS five_prime_UTR splice and three_prime_UTR are already included in the exons Remove those three by changing their priority to 0 double click on the 1 to edit it and click Next Specify an output file e g Rcount_rice_annotation xml in your working directory and click Next Finally close the program by clicking on Finish 2 6 Counting the number of hits per gene with Rcount distribute Use Rcount distribute to count the number of reads per gene Open the program and click new to create a new project This will open a multi tabbed dialog For each sample specify following files under the Files tab Input gt Annotation Rcount_rice_annotation xml stored in path to rice tutorial Input gt Alignments accepted_hits_weighted bam stored in path to rice_tutorial SRR976339 Output gt Alignments accepted hits weighted mapped bam to be stored in path to rice_tutorial SRR976339 Output gt Counts SRR976339_counts txt to be stored in path to rice tutorial Note that the example refers to the sample SRR976339 The bam files can be found and should be stored in the corresponding folder path to rice_tutor
20. tables starting from initial read files The example data is from O sativa and comprises two sperm cell samples 1 Download and unpack the archive rice_tutorial zip from www botinst uzh ch research development grossnik rcount html It con tains a folder with the rice reference genome and its annotation in gff format MSU7 from rice plantbiology msu edu It additionally contains pre processed bam and bai files in case you would like to try only Rcount and to skip the download and alignment part of the tutorial in this case go to section 2 4 The short reads download and alignment part is written for an Ubuntu like Linux 2 1 OPTIONAL Installation of additional programs Additional programs are required to download and align the short reads It is later assumed that these programs reside in a folder that is included in your PATH environment variable This can be done by either moving the programs into one of the by default included folders e g usr local bin or by adding the folder containing the programs to the PATH environment variable Note that the latter is a temporary solution the commands have to be entered each time you start a new terminal Code for both options is given for each of the programs note that the hash tag stands for comments which do not have to be typed into the terminal e SRA toolkit Visit www ncbi nlm nih gov Traces sra sra cgi view software down load the archive for Ubuntu Linux 64 bit arch
21. tez R Marin D Toledo Flores L Froidevaux A Liechti P Wa ters F Gruetzner H Kaessmann Origins and functional evolution of Y chromosomes across mammals Nature 508 7497 2014 488 493 17
22. therefore not necessary e The NH i x tag must be present Without this tag multireads are not recognized by Rcount multireads and treated as unique alignments e There should not be any XW or XM tag Recount uses XW f x to store the weights of reads with multiple alignments and XM i x for mapping statistics The latter is similar to the FLAG column of a bam file The individual bits are explained in table 2 e The CIGAR string should only contain the operations M N I D Other operations are not recognized by Rcount distribute Note regarding paired end reads Rcount distribute automatically takes the first read of one pair and ignores the second Table 2 Description of the individual bits in the XM i x tag introduced by Rcount distribute bit description of the alignment 0x1 belongs to a multiread 0x2 has been skipped by Rcount distribute 0x4 contains gaps 0x8 has a weight of zero 0x10 maps to a known locus 0x20 maps to a known exon 0x40 maps to a known splice junction 0x80 maps to ambiguously several loci 3 2 Rcount format Creating the genome annotation file To overcome the large variety of gnome annotation formats and their sometimes loose definition Rcount uses a novel format which follows clear structural rules while still offering flexibility to add new features The conversion of the most common formats gtf gff into the xml format is done by Rcount format Aside the format conversion it also enables t
23. tion structure their priorities are only compared between each other The pseudogene would not be compared to the gene s sub feature mRNA Likewise an mRNA is not compared to an exon e Features must have three levels Rcount distribute assumes three annotational levels for example gene mRNA exon Features without three levels are ignored For convenience several pre processed genome annotations A thaliana B tau rus C elegans C familiaris D melanogaster G gallus H sapiens M muscu lus and R norvegicus are provided on www botinst uzh ch research development grossnik rcount html in the archive test_data_annotations zip They were build with the data available on tophat cbcb umd edu igenomes shtml using the ENSEMBL data source The archives on this website also contain pre built bowtie and bwa indices which can directly be used in conjunction with the provided genome an notation To create the individual xml genome annotations the genes gtf files were first pre processed with the python script convertCufflinksGTFforRcount py supplied in the archives linux 64bit zip windows 64bit zip and mac 64bit zip and then processed using Rcount format with the ENSEMBL option enabled Each of the genome annotation has been tested with a random sample from SRA see section 4 for more details 3 3 Rcount multireads Weighting reads with more than one alignment Some organisms have very large gene
24. tribute unique multiple time memory time memory A thaliana 41 098 964 2231 636 12 min 1 0 Gb 18 min 1 Gb B taurus 34913558 3538481 15 min 1 1 Gb 18min 928 Mb C elegans 25492284 178897021 10 min 729Mb 14min 1 0 Gb C familiaris 66816362 4809058 18 min 2 8 Gb 22min 878 Mb D melanogaster 25 806 514 969484 9min 216 Mb 12min 981 Mb G gallus 10 465 361 158067 2 min 185 Mb 4min 803 Mb H sapiens 16056845 2169498 5 min 1 0 Gb 13 min 3 3 Gb M musculus 9 630 768 7898165 8 min 3 2 Gb 8 min 1 7 Gb R norvegicus 21729100 1288736 10 min 410 Mb 12 min 1 2 Gb 14 5 Building from source If the binaries are not working or you would like to implement a new feature you can build them from the source code 5 1 Linux Compiling the programs using QtCreator is quite easy 5 2 download and unpack the archive source zip install QtCreator on the Ubuntu repository qtcreator make sure to use Qt 4 x x libraries install zlib on the Ubuntu repository zlib1g zlibig dev zlibig dev which is required by Rcount multireads and Rcount distribute start QtCreator and open each of the pro files Rcount format source p502_format_wizard p502_format_wizard pro Rcount multireads source p502 process multireads p502 process multireads pro Rcount distribute source p502dataAnalysisRNA p502dataAnalysisRNA pro for Rcount multireads and Rcount distribute copy the seqan folder into one of your general include
Download Pdf Manuals
Related Search