bretonics / chomp Goto Github PK

Possibly switch to STAR RNA-seq aligner to search sequence to CRISPR sequence matches.

Search CRISPR targets in reverse input sequence.

Get match hits for each CRISPR and rank matches according to how many bp are matched. Not only how many occurrences per target throughout genome but which has the least matching base pairs.

Fix output directory to a be under default top level gRNA directory

gRNA
    |_gene1
    |_gene2
    |_gene3

The number of off-target sites should be minimized, thus the number of occurrences (number of times the CRISPR oligo matches back to the sequence -- may it be genomic or onto itself--) the least number of off-target matches the better.

The number of occurrences per CRISPR query match is incorrect. Each successive match gets it's previous match added to the number of occurrences.

This array is not flushing.

Seems to work without -genomes option...what don't I see

Subroutines:

• setParameters (manage #3)
• getSequence subroutine to return sequence details with BioPerl

In cases where crispr sequences are being searched in sequence-seq, where the same gene sequence is present in the genome -genome, output should reflect handling exact (23 bp) matches.

Since same gene sequence is searched for crisprs, there will be exact matches in the genome for each crispr sequence found. We want to base sorting on 2nd most matched bp in this case and omit exact matches.

Handling of sequence information should reside in module and check for errors:

• mv getSeqDetails to Search module
• croak if sequence empty

Report should have headers with descriptive comments on how the results were generated.

ADD

• Add call parameters
• Time/date
• Number of hits per subject
• Total number of (no)/hits

Add support to BLAST against multiple chromosomes or sequences.

• append name (-out and subject sequence) to blast.html output. Ex.) blast_geneX_ChrX.html
• mv output to CRISPRS/blast/
• support directory vs. files

When a crispr sequence has no matching sequences in search sequence, then nothing is stored in hash. This causes an "undefined value" error when sorting unique identities for each match.

• The integrity of the 'targets' hash must be checked to include all crispr occurrences (keys)
• If no matches found for crispr, then store appropriate info in attributes

Anonymous hash data structure incorrectly constructed at line 54.

May be related to #6 , but needs to correctly pair key->value

'CRISPR_9' => [
                          {
                            'qseqid' => '1',
                            'qend' => '25',
                            'qstart' => '23',
                            'nident' => undef,
                            'pident' => '23
',
                            'sstart' => '47',
                            'send' => 'plus',
                            'sstrand' => '100.000'
                          },
                          {
                            'pident' => '15
',
                            'sstart' => '213',
                            'send' => 'plus',
                            'sstrand' => '93.750',
                            'nident' => undef,
                            'qseqid' => '1',
                            'qstart' => '16',
                            'qend' => '198'
                          },

Add support for ViennaRNA secondary structure prediction.

Allow multi fasta file for -genome option.

Currently searches against multiple files passed in -genome option, but only searches first sequence in multi fasta file (or first '>').

Allow the sorting to be increasing or decreasing of occurrences/identities. This can be achieved in 3 different ways:

• -sorting option setting each or both simultaneously
• -occ option to allow sorting by occurrences to be inc/dec
• -ident option to allow sorting by identities to be inc/dec

After fixing issue #20, the removal of CRISPRs with 'No Hits' results in removing them completely when sorting on line 337.

#20 fixes 'undef' issue, but introduces bug where CRISPR may not necessarily be missing in all subject sequences (not only from the singular subject where it is not found). Currently it completely omits those others.

Searching CRISPR targets against subject_1
CRISPR_22 has 0 hits here
CRISPR_50 has 0 hits here
CRISPR_72 has 0 hits here
CRISPR_93 has 0 hits here
CRISPR_129 has 0 hits here

Searching CRISPR targets against subject_2
CRISPR_25 has 0 hits here
CRISPR_91 has 0 hits here

So instead of omitting CRISPRS 22, 50, 72, 93, 129 only for subject_1, it does it for all subjects.

If 'No Hits' CRISPRs are not removed, sorting fails. Conundrum.

Different CRISPR sequences in BLAST results than from 'crisprs.txt' file containing detailed information.

• ✔︎ FASTA file names matche with HTML output
• ❌ BLAST output does not match 'crisprs.txt'

Add the following parameters as command line arguments:

• -genome: Genome file to use in blast search instead of self -seq option