[asurarocks]

Also, If you are looking for analyzing MeDIP-seq data that has replicates then you should probably try MeDUSA. A pipeline that uses MEDIPS, USeq and DESeq for finding and annotating DMRs.

[fczqx]

Hi asurarocks,
thank you for your nice help, but when I run LoadRatsGFFApplication, it always throw errors, I think it seems due to missed to load database of expriment data. But I sequenced the samlple by MeDIP-Seq.I don't have such format data to load into table of medip_expt. Would you mind send me the command of LoadRatsGFF?

Also, If you are looking for analyzing MeDIP-seq data that has replicates then you should probably try MeDUSA. A pipeline that uses MEDIPS, USeq and DESeq for finding and annotating DMRs.

[asurarocks]

Hi fczqx,

for LoadRatsGFF I used the following syntax

java batman.LoadRatsGFFApplication -expt EXPT -dbURL jdbc:mysql://domain.edu/dbname -dbUser username -dbPass password FILE-chr2.pad.gff

I hope it helps.

[fczqx]

Hi asurarocks,
Thank you for your help, it works now. But I still hava an question is about the pad.gff file, it shows the format as follows:
MT ReadsToPAD probe 151 151 9.0 . 0 probe "CHRMT_0000000151" ;ROI "CHRMT"
The value of column 4th and 5th are always same. Does the first 151 means the start position and the second 151 means end position? If yes, I think my test data may take something wrong. By the way, would you mind send your email address to me? You can email me as: [email protected]. Thank you again for your help.

[asurarocks]

Hi fczqx, I think the format is correct. I no longer use batman since I switched to MeDIPS (and then to MeDUSA) for my analysis.

[fczqx]

Hi asurarocks,
If the format is correct, I think I should check previous steps, because when I Calibrate the genome, it report error with" Baseline=NAN Response=NAN", I try read the source code of Calibrate, It seems say that the coupling profile overflow the boundary. Have you met this situation ever before or what may caused by that? If you have, can you help me again to sovle it? I think the final result is almost there. I'll really appreciate it!
My type the Calibrate as following:
java -server batman.CalibrateApplication -dbURL jdbc:mysql://domain.edu/dbname -dbUser username -dbPass password -couplingProfile ms -expt medipseq -chrMT -writeDB.

[fczqx]

Hi asurarocks,
I figured out the error of Calibrate, but there has an new error which is "rejecting due to boundary". It seems that the query position is less than 1 or more than the chromosome's length....Do you know way?

[fczqx]

Hi asurarocks,
do you know how to set the parameter of "precision" with SampleMethStatesApplication? It seems should be estimated by ourselves and will be changed with different samples? Thank you!

[fczqx]

And if we set, which threshould value may be the best?

[asurarocks]

Hi fczqz, Sorry, I am not sure about those things. Batman takes unusually long time writing/reading database, it became unfeasible for us to use it for any of our analysis. So I don't have much experience with it other than the steps I pasted before. Maybe you can write to Thomas directly about your questions (he is quick in replying emails).

[fczqx]

Hey asurarocks,
I switch to use MEDIPS instead of batman... Cause batman is too slow to calcuate in my computer...and MEDIPS seems run much fast than batman. Thank for your advise but now I have another question is when I compair with two MEDIP samples, the MEDIPS program seems need us to input a INPUT.SET file as ground file to calcuate p-value? Does this INPUT file is really required? And what does this file maining stand for? Thank you advance.

[asurarocks]

Input file is non-enriched seq experiment, serves as a background to measure relative increase or decrease in methylation. It is totally optional. You can just go ahead with your control and treatment, without worrying about the input.set

[sciencewu]

MEDIPS is a friendly package for analysis the MeDIP data, maybe you can try it.