TO_masterpipe_onNewClus.pl

TarO_indevel.pm

Accepted command line parameters:

-i [Inputfile (must give full path name)]
-o [output directory (must give full path name)]

Optional input:
-s [integer] specifies how many sequences per file when splitting (input) fasta file to run jobs on the cluster
-t [local cluster node directory] specifies location of local temporary storage
-p [directory] location of perlscripts to run jobs on the cluster
-j [directory] location (top of the tree) for storage and reading of raw output from pipeline steps
-d debug
-h print this message
-v verbose mode
-z overwrite all cluster-generated results
-q minimal upload (only upload data into new display tables - used for incorporating legacy queries (v0.1) into v1.0)
-b [integer] specifies the maximum number of sequences to include in the multiple sequence alignment
-w [local cluster node directory] location for local psiblast database
-a [directory] location of databases and .dat files
-m [stepname(s)] overwrite specific stepname(s)
Available stepnames:
disembl
globplot
targetdb
pfam_rps
cdd_rps
smart_rps
kog_rps
cog_rps
signalp
tmhmm
muscle
ronn
jpred
netnglyc
netnglyc
netphos
pdbpsi
obscore
parcrys

Below are the comments from the script

1. Deal with legacy queries.
2. Read input into hash Explicitly mark sequences as user input (to avoid id duplication problems)
3. Initialize the script
4. Do some file manipulation
5. Check for results of previous runs to avoid rerunning analisys
6. Do the database update if necessary
7. BLAST sequences vs COG database (update status to Completed)
8. Analyse the BLAST results to identify sequences  with sufficient identity to be included into a COG (TODO - use Rost Thresholds)
9. Read in all members of that COG Parse COG information groupings
10. Read in the cog data including phylogenetic information & full organism name
11. Read in protein sequences
12. OUTPUT COG orthologues and User-input sequences to a single file as input for PSIBLAST SEARCHING
13. ### Write out files for sequences with known Gram+/- and euk origin (archaea currently processed as eukarya by signalP) + files for unknown remainder
14. PSIBLAST the whole COG (plus original input sequences) against UNIREF100  HERE! this will be transformed into a lookup once the allvsall has finished....? Or perhaps do a profile-profile search? (update status to Completed)
15. Parse psiblast output (Finding UNIREF (PSIBLAST) homologues)
16. Output homologues (currently just from psiblast of uniref100 (UNIREF homologs)
17. Concatenate all homologues & orthologues plus original inputs into single file (Strategy to remove redundancy to be implemented.... here!!)
18. Analyse the inputs/orthologs/homologs to remove redundancy (remove sequences (100%coverage + 100%identity criteria)) plus write out file with orthologues + X number of homologues to make multiple alignment from (X is possibly user defined variable describing the maximum number of sequences to be included in the alignment  (caveat that all orthologues will always be included)) 
19. Draw together all sequences into a single object & assign UniProt equivalence to each sequence
20. Assigning UNIREF length & organism info.
21. Assign Phylogenetic information to UniRef100 + input sequences
22. Write out files for sequences with known Gram+/- and euk origin (archaea currently processed as eukarya by signalP) + files for unknown remainder (Writing out signalP phyla files)
23. Run Muscle to generate multiple alignments for top 200 seqs: (+ add-on POA later??) # Strategy allows the alignment of a maximum of 200 sequences 
    #- including orthologues and top-hitting homologues
    # COG orthologues are given priority,
    # sequences are ordered by evalue match to the input sequence(s)
    # sequences more than 1000aa longer than the input sequence are excluded
24. Order the cog orthologues according to their psiblast evalue match & exclude any seqs more than 125% of longest input seq (only take sequences up to 1/4th longer that the user input sequence)
    
    UP UNTIL THIS POINT ALL ANALYSIS ARE RUN UNCONDITIONALLY.
    ANALYSIS OF SEQUENCES PIPELINE (start if close)
    
    
25. Protein disorder prediction for now just use Globplot and Disembl. Disembl use homologues plus orthologues now. No trimming of orthologues now conducted: just downweight in the ranking scheme
26. Running RONN
27. Running Gloplot
28. Running BLAST TargetDB
29. Running PSIBLAST PDB (only input sequence)
30. Running BLAST PDB - FOR ORTHOLOGUES & HOMOLOGUES
31. Search CDD (RPSBLAST - Pfam)
32. Running RPSBLAST CDD
33. Running RPSBLAST Cog
34. Running RPSBLAST SMART
35. Running RPSBLAST KOG
36. Run ParCrys (& calculate OB-scores as part of this process)
37. Running TMHMM2
38. Glycosylation site prediction NetNglyc
39. NetOglyc
40. Phosphorylation site prediction Netphos
41.  Signal peptide prediction (signalP) - eukaryotic seqs
    gram+ve seqs, gram-ve seqs
42. JPRED - run only on input seq(s) (end of condition)
43. Miscellaneous analyses (count methionines, cysteines etc., calculate PI, Mr, length, hydrophobicity
44. Calculate extinction coefficient at 280nm (molar and 1mg per ml coefficients)
45. Get Nucleotide Sequence....  Nucl. seq dependant annotation:  GC content,Tev protease sites and codon usage relative to e.coli/dictostelium/insect cell/cell free expression.
46. Reading targetDB blast and PDB psiblast info to Identify psiblast queries with possible similar structure to a db sequence (using combos of id, alignlen, & evalue)
47. Read in Protein disorder prediction results (to output to the tab delimeted file for uploading into the database)
48. Read in Parcrys info
49. Read in SignalP results
50. Read in TMHMM2 results
51. Read in Jpred results
52. Read in netNglyc results
53. Read in netPhos results
54. Make MSA annotated with Globplot/Disembl disorder predictions (for Jalview)
55. Copy the groups file, MSA and jnet file to be accessible in the website Jalview directory
56. Write out tab delimited files for the database
57. SEQUENCE RANKING MECHANISMS Currently ParCrys Score (& blast evalue) additions to sequence table
58. Steps to deal with legacy data - to add in the extra data found for the given query
59. UPDATE STATUS TABLE
60. email user to indicate job is ready
    

• 137285 (bytes)-TarO_indevel.pm (main script library)
• 31727 -Clust_int_taro.pm (methods to send jobs to the cluster)
• 22831 -Blast_pars.pm (BLAST parsers)
• 21180 -ParCrys.pm (methods to calculate ParCrys scores)
• 17795 -BioSession.pm (HTTP session management)
• 13422 -OB.pm (methods to calculate OB score)

• 63997 - TarO_indevel.pl (most of the modules above are predominantly called from this script)
• 3188 - ParCrys.pl
• <3000 -15 other scripts (most are the data parsers)