Re: Is there any way to display the highest possible score and relative score for a TFBS

[Cassandre Coles]

Hello,

I think I figured out how to calculate the relative score and PWM (which seems to be the score). 1) I wanted to know if W is the same as the score for each site? Would the Sj be considered 1 for the residue in the sequence and all others are 0?

I am really asking if you would be able to figure out why my actual score doesn't match the results I got from JASPAR? Also, would you suggest I calculate this differently? I attached my calculations in this email. Thank you so much for your help!

Stay Safe,

Cassandre

On Wed, Apr 8, 2020 at 4:48 PM Cassandre Coles <cco...@uic.edu> wrote:

Hello,

I scanned a murine genomic sequence for an OTX2 binding site.

I have several questions.
1) What is threshold?

I lowered the threshold to 70% and I got 14 hits. Threshold means it eliminate false positive right? So the more I lower this value the more false positives I get? And these sites more than like may bibd in vitro but may have no function?

2) What does score and relative score mean?

I saw the equations but it lead to more confusion. I really want to know if a score of ~7 snd relative scie atound 0.8 is acceptable to consider a sequence a potential TFBS.

The column title score and relative score are a bit confusing to me. What are the highest possible scores and relative scores? Would it be possible for those score to be calculated and displayed for folks like me? I have read the papers you have suggested on your site and I've went to similar questions but they dont really make everything make sense for me. I want to know what does relative score actually means. I think I know what score is to but I would love clarification for that as well. Bioinformatics is not my field.


Thank you so much!

--
You received this message because you are subscribed to a topic in the Google Groups "JASPAR Q&A Forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/jaspar/F7D9yR4U79w/unsubscribe.
To unsubscribe from this group and all its topics, send an email to jaspar+un...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/jaspar/c67cdcba-97d2-486e-b36d-3ee6f6c4432c%40googlegroups.com.

[Oriol Fornés]

Hi Cassandre,

I highly recommend that you follow the instructions posted in our GitHub repository:

https://github.com/wassermanlab/JASPAR-UCSC-tracks

It should be as simple as:

./scan_sequence.py --fasta-file $FASTA_FILE --profiles-dir ./profiles/ --output-dir ./Otx2/ --profile MA0712.2

Where $FASTA_FILE contains the sites that you want to scan in the FASTA format:

>SITE_1

SEQUENCE_OF_SITE_1

>SITE_2

SEQUENCE_OF_SITE_2

...

>SITE_N

SEQUENCE_OF_SITE_N

Results will be placed inside the folder ./Otx2/ as a tabulated file in the following format:

site_number[TAB]start[TAB]end[TAB]OTX2[TAB]relative_score[TAB]p-value[TAB]strand 

Happy Easter!

To unsubscribe from this group and all its topics, send an email to jaspar+unsubscribe@googlegroups.com.

[Cassandre Coles]

Hello,

I think I found what you described but I was wondering if you could give more detailed instructions. Could I get the order of what I should click to get to where I need to be? I am a bit lost. Thank you so much!

[Oriol Fornés]

Hi Cassandre,

If you have little bioinformatics experience, here is another way of doing it:

1) Go to the latest OTX2 profile from the JASPAR CORE (i.e. http://jaspar.genereg.net/matrix/MA0712.2);

2) Add the OTX2 profile to your cart; and

3) Go to your cart;

4) Insert your sequences (in FASTA format) in the text box and scan these sequences with the OTX2 profile.

You should obtain something like this:

Hope this helps!

--

Oriol Fornés Crespo, PhD
Postdoctoral Fellow @ Wasserman Lab
Centre for Molecular Medicine and Therapeutics (CMMT)
University of British Columbia

BC Children's Hospital Research Institute 
950 W 28th Ave, Room 3109, Vancouver, BC V5Z 4H4, Canada

E-mail: or...@cmmt.ubc.ca

On Fri, 10 Apr 2020 at 06:18, Cassandre Coles <cco...@uic.edu> wrote:

Hello,

I think I found what you described but I was wondering if you could give more detailed instructions. Could I get the order of what I should click to get to where I need to be? I am a bit lost. Thank you so much!

--
You received this message because you are subscribed to the Google Groups "JASPAR Q&A Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to jaspar+un...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/jaspar/e4dfc507-c8e6-4869-8bdd-0d4430b3f410%40googlegroups.com.

[Cassandre Coles]

Hello Oriol,

Thank you for your patience! 

I tried it both ways. First, I downloaded Python 3.8.2. and Github on my machine and then I had someone walk me through it. I believe I followed the instruction correctly but it didn't work out. I saw that you used model MA0712.2 and I am using PH0130.1. For now I am interested in scanning the murine genomic DNA for the OTX2 binding site. I highlighted what I am talking about.  /scan_sequence.py --fasta-file $FASTA_FILE --profiles-dir ./profiles/ --output-dir ./Otx2/ --profile MA0712.2  Will that impact my results? 

For the second method, I believe we are on the same page. I used JASPAR to scan my murine genomic DNA sequence and I got 17 hits. I got the same results as before which is nice.

 I really started all of this by asking what does score and relative score mean? Am I correct in thinking that the higher the relative score the higher the chance of the TFBS being a functional binding site? It seems like I am using score to calculate relative score, which seems to be more useful for indicating whether or not a site is worth studying. 

I am hypothesizing that the increase in murine OTX2 and the increase in a gene may be linked. I justify linking these two using the data I have found here. I believe the first hit (and maybe a couple below) may be an actual candidate, based on the relative score. Thank you for being patient with me! I appreciate all your help.

Best,

Cassandre

To unsubscribe from this group and stop receiving emails from it, send an email to jas...@googlegroups.com.

[Oriol Fornés]

Hi Cassandre,

The profile that you are using to scan your sequences, http://jaspar.genereg.net/matrix/PH0130.1/, was derived from protein binding microarray (PBM) data in 2009.

The profile from the CORE that I am suggesting, http://jaspar.genereg.net/matrix/MA0712.2/, was derived from >100,000 human ChIP-seq peaks in 2018.

PBM is an in vitro assay. In contrast, ChIP-seq is an in vivo assay, and is expected to capture better how transcription factors bind to DNA in the genome.

Moreover, the DNA-binding domains of the human and mouse OTX2 factors have the exact same protein sequence. Thus, one can expect that they will bind the exact same sequences in vivo; the human profile can be used to scan the mouse genome and vice versa.

Hope this helps!

On Thursday, 9 April 2020 14:10:44 UTC-7, Cassandre Coles wrote:

To unsubscribe from this group and all its topics, send an email to jaspar+unsubscribe@googlegroups.com.

[Muhammad Danial]

Hello everyone I have an assignment :::: Download PWM of CTCF from Jaspar.
 I read the documentation but still, I don't understand how to download the PWM of Ctcf, please can someone help me how can to download the PWM.
Thank you

To unsubscribe from this group and all its topics, send an email to jaspar+un...@googlegroups.com.

[Anthony Mathelier]

Hi Danial,

We have 3 PFMs from CTCF: https://jaspar.genereg.net/matrix/MA0139.1/, https://jaspar.genereg.net/matrix/MA1929.1/, and https://jaspar.genereg.net/matrix/MA1930.1/. You can go on the corresponding webpages and directly download the matrices.

Best