how to identify a 14-20 nt specific for chromosoe #1 which have the maximum recognation sites for Chr. #1

Question

[Dan]

How to identify a 14-20 nt specific for chromosoe #1 which have the maximum recognation sites for Chr. #1 .

The linux command would be preferable but any software suggestions and kind help really appreciated.

Best,

Dan

[GenoMax]

Which genome? Doubt one can find a sequence like that in terms of the size specified, specificity (only for chr 1) and with maximum recognition sites.
Any additional information you can provide to clarify the request?

[Dan]

Hi,

Thank you very much for your prompt reply. Chr#1 was just an example. I more interested in X and Y Chromosome to do FISH. right now the FISH probe that I am using and I received from someone who dose not want to reveal the sequence to me is not specific enough and has low intensity.
My intention is to find a/sets of 14-20 nt sequence with high X/Y chr. recognition site(s) to i) get specific signal and ii) higher intensity due to higher number of recognition sites (repetitions in X/Y Chr.)

[GenoMax]

That is useful information. Someone else may need to help you further. Perhaps this site, and this software may help while you wait to hear from experts.

[GenoMax]

Some guidelines for FISH probe design (from http://www.exiqon.com/custom-fish). These may be universally applicable.

Detection probes are typically 20-25 nucleotides in length. However, shorter or longer probes can also be used.
Avoid stretches of 3 or more Gs or Cs.
Avoid stretches of more than 4 LNA™ bases, except when very short (9-10 nt) oligonucleotides are designed.
Avoid LNA™ self-complementarity. LNA™ hybridizes very tightly to other LNA™ residues.
Keep the GC-content between 30-60 %.
A Tm of approximately 75 °C is recommended.
No LNA™ bases should be placed in palindromes (G-C base pairs are more critical than A-T base pairs).

[Dan]

Thank you very much for your kind helps.

Best

Answer 1

[5heikki]

1. Break the chromosomes into k-mers of desired length (e.g. with jellyfish)
2. Combine non-target k-mers into one file
3. Sort target and non-target k-mer files
4. Use comm to find k-mers unique to target chromosome
5. From the unique k-mers select the one with the highest count

If you run out of memory, skip step 2 and instead do the comparisons one by one always using the remaining unique k-mers as the other input file for comm.

Answer 2

[Pierre Lindenbaum]

brute force, not fully tested: scanning all the kmers in a fasta stream.

g++ -O3 -Wall biostars201509.cpp 
./a.out < human_g1k_v37.fasta
(...)
CTGGACATAACA

Answer 3

[Dan]

Thank you so much Pierre Lindenbaum. You are awesome! Best regards,

Dan