SOLiD quality data

[mbm]

We recently completed a ChIP-seq run on the ABI SOLiD. We received the following data back for one of our samples. These data were from a quarter (1/4) of a slide.

------------------
61993194 total beads found
Total Beads
0 mismatches 5631945 ( 9.08%)
1 mismatches 5976039 ( 9.64%) 11607984 ( 18.72%)
2 mismatches 5155084 ( 8.32%) 16763068 ( 27.04%)
3 mismatches 4243237 ( 6.84%) 21006305 ( 33.88%)

Uniquely Placed Beads
0 mismatches 3957369 ( 6.38%)
1 mismatches 4341342 ( 7.00%) 8298711 ( 13.39%)
2 mismatches 3894787 ( 6.28%) 12193498 ( 19.67%)
3 mismatches 3351175 ( 5.41%) 15544673 ( 25.07%)

Errors within Uniquely Placed Tags
Total Errors 22184388
Single Errors 19528902 (88.03% of Total)
Adjacent Errors 1355877 (12.22% of total)
Valid 716317 (6.46% of Total) (52.83% of Adjacent Errors)
Invalid 639560 (5.77% of Total) (47.17% of Adjacent Errors)

Starting Points within Uniquely Placed Tags
Number of Starting Points 15186453
Average Number of Reads per Start Point 1.02

Starting Points within Uniquely and Randomly Placed Tags
Number of Starting Points 19891182
Average Number of Reads per Start Point 1.06

Coverage of Uniquely Placed Tags
Bases Not Covered 2643970633(85.08%)
---------------

Do these numbers look promising to others as far as sequences covered? Are these similar to other people? The samples were human and were mapped to HG18 reference genome. The unique placement of sequences seems low to me from the data ABI was quoting (5mil total). I apologize for lack of detail, however I want to get opinions without biasing people.

Thanks!

[Chipper]

It looks like you have 15mil uniqely aligned reads, not 5 mil? I think that is a very good number for 1/4 of a slide. Is this 35 bp reads? If it is longer you may need to allow for more mismatches.

[mbm]

yes, sorry, 15 million, I forgot the 1 in front of the 5! ABI was projecting >20 mil. They are 35bp reads. Thanks for the input, just want to know how our run stacked up with others.

[Chipper]

That is strange, are you sure this was 1/4 only? Send a pm if you want to discuss details.

[ECO]

62 Million beads for a quad is on the very high end of normal. The manual officially recommends 30M.

You can recover quite a few more mapping reads by filtering as suggesting in Cloonan, et al...iteratively filtering and trimming.

[mbm]

Thanks for the Cloonan tip. I read the paper but must have glazed over the very first sup fig. we will try that and many other things.

Have you noticed a problem with loss of GC rich regions? When the P1 linkers are added and amplified (before bead attachment and emulsion PCR), we believe GC rich regions aren't be amplified or they are not being sequenced. A control gene is not showing amplification directly at the promoter, a site of binding known for 15 years and assayed by ChIP and ChIP-chip quite easily. Had a similar experience with Sigma's WGA kit. Any thoughts on that?

[yul2]

Hi mbm, now I am ready to do CHIP on SOLiD system .while see you also doing CHIP application ,i am very happy. I have some inquiries about CHIP test. The first one is how to increase the amount of DNA , because the amount is low ,maybe only 10-30 ng. Could you send me your protocal about sample preparation and construct the fragment library?

[mbm]

yul2,

I did not personally create the fragment library. We collaborated with Agencourt for our library prep and sequencing. However, the samples that I sent them were not "amplified", I instead collected ChIP samples in triplicate and combined them to reach a total sample amount of ~400ng. This was enough for them to attach the P1 linker and run through a LMPCR amp prior to bead attachment and emulsion PCR. These sequences and protocols are available from ABI. Although the sample amount was 400ng, I believe that this should be able to completed with far less sample than that in the future. This was before the version 2.0 chemistry and methods. We are going to try a library prep or amplification ourselves sometime in the near future.

[subh]

Please tell me how can select best quality value for SOLiD results. what is the criteria for selecting quality value??

Plz help. I am stuck with results.

[SeqAA]

wow. nice thread necro. Are you trying to filter by QV?
I usually use the QV10 to trim away some reads.