Hot to calculate coverage on a 454 targeted region sequenced ?

[Giorgio C]

Hi all,

I have a region of 250.000 bp targeted with probes and then sequenced on a 454.

How can i make a statistical calculation of the coverage ?
Do i have to consider the total number of bp sequenced and number of reads ?

Or just for example considering 1.000.000 bp sequenced :

Coverage = 100.000/250.000 bp ?

Any advices are highly appreciated,
Thanks

[sphil]

Hi,


coverage = (#reads * length of reads) / length of target sequence

thus, having 50k reads with length of 250 and 100k target sequence length you get a average coverage of 125. However please keep in mind that some regions can be sequences more easily and all sequences techniques have their own caveats. Keeping that in mind the above approximation is still valid.

Best

Phil

[Giorgio C]

Thank you very much for your reply. So if i undersand well the correct formula is:

total numbero of bp obteined * avarege length of all reads / length of targeted region

Is that right ?

[Giorgio C]

or is :

total number of bp obteined * total number of reads / length of targeted region ?

[sphil]

it is: total number of bp obtained / length of target region. Since total number of bp obtained == amount of reads * length of reads.

[Giorgio C]

Hmm.. i'm sorry but something is not clear to me; If i have this situation:

- Number of total Bp sequenced = 7.114.424

- Number of reads = 354

- Length of targeted region = 225.000 bp


Is correct only to do 7.114.424 / 225.000 = 31,61 ; Means 31% of coverage ?

And i don't need to consider the "Number of reads" (354) ?

I hope I was clear,
Thanks you for your help

[sphil]

Hmm.. i'm sorry but something is not clear to me; If i have this situation:

- Number of total Bp sequenced = 7.114.424

- Number of reads = 354

- Length of targeted region = 225.000 bp


Is correct only to do 7.114.424 / 225.000 = 31,61 ; Means 31% of coverage ?

And i don't need to consider the "Number of reads" (354) ?

I hope I was clear,
Thanks you for your help

yes, you are right. You do not need to consider the number of redas because this should result from: number of reads * read length. To get the coverage you simply do the calculation above, thus 7.114.424 / 225.000 = 31,61. That does not mean that you ave 31% coverage. It means that every base in your target sequence is covered by 31 bases contained in your reads. So you have a coverage of 31, which is not too bad.

[Giorgio C]

Thank you very much for your reply. You have been very helpful !!!

[jordipt]

Hi there,
I know it is quite late but I think it might also be interesting to others to leave a comment on this (still open) thread.
When considering coverage of a targetted run you should also include in the equation the % of "on target" reads. That is, in a typical capture experiment not all reads obtained from the sequencer map to your region of interest, in fact, for a region like the aprox 250Kb you mention, on target reads could be about 70-50% of all your reads. Thus the number of bases obtained that you have to use in your calculations should be all those that map to your target region.
Best

[Giorgio C]

Thanks jordipt for your clarification...it's never late for good suggestions