Skip to content

Latest commit

 

History

History
114 lines (76 loc) · 6.22 KB

about_cfg.md

File metadata and controls

114 lines (76 loc) · 6.22 KB

CONFIGURATION FILE

0. NORMALIZATION and DATA TYPE

A. Normalization Type

PORT offers Exon-Intron-Junction level normalization and Gene level normalization. Select the normalization type by setting GENE_NORM and/or EXON_INTRON_JUNCTION_NORM to TRUE. At least one normalization type needs to be used.

B. Data Type

i. STRANDED

Set STRANDED to TRUE if the data are stranded.

ii. FWD or REV

If STRANDED is set to TRUE, strand information needs to be provided. Set FWD to TRUE if forward read is in the same orientation as the transcripts/genes (sense) and set REV to TRUE if reverse read is in the same orientation as the transcripts/genes (sense).

Note: when dUTP-based protocol (e.g. Illumina TruSeq stranded protocol) is used, strand information comes from reverse read.

C. Chromosome Names

🔴 By default, PORT ONLY uses numbered, X or Y (e.g. chr1,chr2,...,chrX,chrY OR 1,2,...,X,Y) as standard chromosome names.

i. File of chromosome names [optional]

Provide a full path to file of chromosomes names (CHRNAMES) if your chromosome names do not follow the chromosome nomenclature described above. For example, for dm6, the CHRNAMES file should look like this (one name per line):

chr3R
chr3L
chr2R
chrX
chr2L
chrY
chr4
chrM

Note: Depending on the genome, it may contain clone contigs that cannot be confidently placed on a specific chromosome (e.g. chrUn) and sequences that is not in a finished state (e.g. chrN_random). We recommend excluding those "non-standard" chromosomes.

ii. Name of mitochondrial chromosome [required]

Provide a name of mitochondrial chromosome (e.g. chrM, M). If there are multiple mitochondrial chromosomes, provide a comma separated list of chromosome names.

1. CLUSTER INFO

If you're using SGE (Sun Grid Engine) or LSF (Load Sharing Facility), simply set the cluster name (SGE_CLUSTER or LSF_CLUSTER) to TRUE. You may edit the queue names and max_jobs.
If not, use OTHER_CLUSTER option and specify the required parameters.

For clusters which only allows job submissions from the head node, specify the job submission host name (HOST_NAME)

2. GENE INFO

Gene information file with required suffixes need to be provided. Gene level normalization requires an ensembl gene info file.

The gene info file must contain column names with these suffixes: name, chrom, strand, txStart, txEnd, exonStarts, exonEnds, name2, geneSymbol, ensemblToGeneName.value.

ensembl gene info files for mm9, mm10, hg19, hg38, dm3 and danRer7 are available in Normalization/norm_scripts directory:

 mm9: /path/to/Normalization/norm_scripts/mm9_ensGenes.txt
 mm10: /path/to/Normalization/norm_scripts/Mus_musculus.GRCm38.84.PORT_geneinfo.txt
 hg19: /path/to/Normalization/norm_scripts/hg19_ensGenes.txt
 hg38: /path/to/Normalization/norm_scripts/Homo_sapiens.GRCh38.84.PORT_geneinfo.txt
 dm3: /path/to/Normalization/norm_scripts/dm3_ensGenes.txt
 danRer7: /path/to/Normalization/norm_scripts/danRer7_ensGenes.txt

Note: PORT uses the mapping of transcripts to genes for gene-level quantification: read counts of all transcripts of a gene are merged for overall gene expression. For annotations with genes with multiple transcripts, make sure the "name" (transcript id) and "name2" (gene_id) are not the same.

Alternatively, you can use a perl script (/path/to/Normalization/norm_scripts/convert_gtf_to_PORT_geneinfo.transcripts.pl) to convert a gtf file to a gene information file.

Note: In this case, check the "gene_id" and "transcript_id" attributes in the gtf file.

3. FA and FAI

[1] genome fasta file

The description line (the header line that begins with ">") MUST begin with chromosome names that match the chromosome names in GENE INFO file(s).

Please check and modify the file appropriately before starting PORT.

ucsc genome FASTA files for mm9, hg19, dm3, and danRer7 are available for download (gunzip after download):

  mm9: wget http://itmat.indexes.s3.amazonaws.com/mm9_genome_one-line-seqs.fa.gz
  hg19: wget http://itmat.indexes.s3.amazonaws.com/hg19_genome_one-line-seqs.fa.gz
  dm3: wget http://itmat.indexes.s3.amazonaws.com/dm3_genome_one-line-seqs.fa.gz
  danRer7: wget http://itmat.indexes.s3.amazonaws.com/danRer7_genome_one-line-seqs.fa.gz

[2] index file

You can get the index file (*.fai) using samtools (samtools faidx <ref.fa>)

[3] samtools

Provide the location of your copy of samtools You can get the most recent source code with git clone git://github.com/samtools/samtools.git

4. rRNA

[1] rRNA_PREFILTERED

Set rRNA_PREFILTERED to TRUE if you prefiltered the ribosomal reads. When rRNA_PREFILTERED is set to TRUE, the BLAST step will be skipped and PORT will not generate percent ribosomal statistics.

[2] rRNA sequence fasta file

rRNA sequence file for Mammal (mm9 - can be used for all mammal), Drosophila melanogaster (dm), Zebrafish (danRer) and C.elegans is available in Normalization/norm_scripts directory:

  Mammal: /path/to/Normalization/norm_scripts/rRNA_mm9.fa
  Drosophila melanogaster: /path/to/Normalization/norm_scripts/rRNA_dm.fa
  Zebrafish: /path/to/Normalization/norm_scripts/rRNA_danRer.fa
  C.elegans: /path/to/Normalization/norm_scripts/rRNA_c.elegans.fa

For other organisms, get rRNA sequences and create a fasta file.

5. DATA VISUALIZATION

Set SAM2COV to TRUE if you want to use sam2cov to generate coverage files. sam2cov only supports reads aligned with RUM, STAR, or GSNAP (set aligner used to TRUE). Make sure you have the latest version of sam2cov. At the moment, sam2cov assumes the strand information (sense) comes from reverse read for stranded data.

6. CLEANUP

By default, CLEANUP step only deletes the intermediate SAM files. Set DELETE_INT_SAM to FALSE if you wish to keep the intermediate SAM files. You can also convert sam files to bam files by setting CONVERT_SAM2BAM to TRUE. Coverage files can be compressed by setting GZIP_COV to TRUE.