SOLiD

[jsun529]

When I run corona_lite matching pipeline on single linux through

submit_scripts_to_PBS.pl -j JOB_LIST.txt

I got these error messages? Any solutions? Thanks.

Parse a job list file containing script names and run them on PBS.

SUBMIT: output_ALIGN_17_1: /root/Desktop/Corona_Lite_Plus_4.2.1/CORONA_LITE/ACADVL.RESULTS/scripts/output_ALIGN_17_1.sh
Can't exec "qsub": No such file or directory at /usr/lib/perl5/5.8.8/IPC/Open3.pm line 246.
open2: exec of qsub -h failed at /root/Desktop/Corona_Lite_Plus_4.2.1/CORONA_LITE//lib/perl/PBSJob.pm line 91
Use of uninitialized value in concatenation (.) or string at /root/Desktop/Corona_Lite_Plus_4.2.1/CORONA_LITE//lib/perl/PBSJob.pm line 102.
[ERROR] qsub: (24768):

[westerman]

My guess, especially considering that you have a "single linux" is that you do not have a batch queuing systems (e.g., PBS) set up. Try just running the commands manually, e.g.,

/root/Desktop/Corona_Lite_Plus_4.2.1/CORONA_LITE/ACADVL.RESULTS/scripts/output_ALIGN_17_1.sh

[jsun529]

Thanks, yeah.
for relative longer insertion, the corona_lite did not pick it up, how to see if there is a good coverage on the known region of the genome with this insertion among the result files corona_lite produces? Thanks a lot.

[westerman]

To check coverage look at the '*.coverage.unique.csv' file.

[jsun529]

Thanks,

For that file I got from test run, it looks like it only have one column with coverage number, how to link back to the reference genome position?

Thanks a lot

[westerman]

Each line is a base. Just cross-reference the line number to the bases of interest. For example if your insert is located in bases 10000 to 12000 in the reference then you should look at times 10,000 through 12,000.

[jsun529]

Great. I guess there is one question immediately follow by that? If there is no good coverage on the known region of the genome with insertion? Does it indicate this is due to the machine did not get the signal(primary analysis), or the software(corona_lite) with the default parameter settings?

Thanks a lot.

[westerman]

In my mind there are 4 possibilities of what is happening.

1) Machine or prep error -- you simply do not have DNA covering the insertion.
2) The sample is missing the insertion
3) The insertion is in a repeat and thus will not be unique
4) The insertion is partially repetitive and corona_lite's parameters miss the insert.

For cases #1 & #2 & #4 you will never (or rarely) find matches. For case #3 you find matches but not in the 'unique' files. Probably the best way to answer that case is to dig into the overall match file and not use the unique match files. The latter are easier to use but the former will give you every place (up to the -z level) where a read matches. Parse this out and do your counts. For example, one of the reads in one of my runs looks like:

>3_53_1869_F3,1_-1030216.3,1_-1030212.3,1_-1030211.3,1_-1030210.3,1_-1030209.3,1_-1030208.3,1_-1030207.3,1_-715310.3,1_-715309.3,1_-715306.3

In other words the read has 10 matches (maximum for my -z of 10) and thus would not show up the 'unique' files. But if I were to parse the above line then I could see that I have matches at positions 1030216, etc.

For case #4 -- bad parameters -- run corona_lite with bigger mismatch limits and see what happens.

It is hard to tell the difference between case #1 and case #2 without re-preping and re-running your sample. Although there are techniques (pcr) which can tell you if your sample actually does have the insert or not without having to do another SOLiD run.

[jsun529]

Thanks, very helpful. What is in average a good acceptable coverage percentage, if applicable.

[westerman]

IMHO the acceptable coverage number is not generally applicable. If you ask ABI they would say that a run with 1 read per every possible base (e.g., start point) is adequate. In other words a coverage of 50 if you are generating 50bp reads. However ABI is thinking of sequencing a single individual to a well characterized genome. If you are doing metagenomics where the number of individuals in a single run is large and where you are looking for the rare SNP then a higher coverage is better.

In your case -- not knowing anything about the size of your organism nor the repeat structure -- I would just take the average non-zero coverage from the unique data and use that number as a way to judge if the non-unique data (and your insert) is being covered at a good level.

We have had a discussion before on SeqAnswers as to if a single number can be used to determine adequate coverage. I am in the camp of using a single number. Certainly when looking the fine structure of a genome there should be different number coverage levels. However I am lazy and just want a single go/no-go number. Your mileage may vary.

[jsun529]

SOLiD

Thanks. how is Corona_Lite comparing to other third party software in analyzing SOLiD data,especially for SNP, indel detections?

Thanks

[mdeng]

My guess, especially considering that you have a "single linux" is that you do not have a batch queuing systems (e.g., PBS) set up. Try just running the commands manually, e.g.,

/root/Desktop/Corona_Lite_Plus_4.2.1/CORONA_LITE/ACADVL.RESULTS/scripts/output_ALIGN_17_1.sh

I had the same Problem. So now I am running the single scripts. Do I have run every script, alone? Or can start e.g. 10 scripts? So, does script1 has to be finished for script2 and so on?

With best from Bonn

//Edit: Ok, I found the dependency entry in the JOB_LIST file. Now all jobs done.

Is there a tool to visualize the Results?