Tweet
Long Mate Pair library construction backgrounder
It is always a gut-check vetting new protocols. Especially ones that take upwards of a week to complete. For me the process is exponentially worse when the protocols offer no rationale for their methods. AB is far from the most villianous in this regard, but it might be worth your time to consider a two issues before diving in to the LMP protocol.
(1) Input DNA quantity and "quality".
I did a stint in an anaerobic protein purification lab early in my career. I was reliably informed by a post doc in this lab that even after a several thousand-fold purification of a protein of interest, to a single band on a stained gel, that every single protein present in the original cells was still in the purified prep. I'm sure this was a little hyperbolic, but the point still stands: purification protocols deplete a sample of the stuff you do not want--they do not completely remove it. Fortunately most of the hitch-hiking biomolecules in a sample DNA prep can be safely ignored. The ones that might trip you up:
RNA. Typically >90% of the nucleotides in a genomic "DNA" prep derive from RNA. (I saw 99.7% less than 1 month ago.) It is usually degraded down to oligos with RNAse so that you may not see it at all on a gel, but it confounds any attempt to measure DNA concentration using UV spectrophotometry. Be particularly suspicious if the sample reads > 1ug/ul concentration (genomic DNA is difficult to dissolve at that high of a concentration) or your 260/280 ratio goes above 1.8.
phenol. Less commonly used these days, but you will still see it. Just a tiny amount will throw your UV spec reading way off! 1 ml of water with 1 ul of phenol dissolved (0.1% v/v) into it will read like 540 ng/ul DNA. But if you look at the full spectrum it is obvious because the peak is at 270 nm.
nucleases. These are a little more problematic. Any self-respecting cell will be replete with all manner of nucleases. DNA preps generally should deplete them down to a tolerable level. But here is the rub: what constitutes "tolerable"? Add a little EDTA (as in TE) around to chelate divalent cations and most DNA nucleases will not function. However, bump the concentration of Mg++ high enough or remove the EDTA and they will be back in action. Since I am going to fragment my DNA using a hydroshear anyway, I don't mind the genomic DNA being a little smeary. But you really want to make sure all the nucleases are gone before you start ligating adaptors on.
(2) Your nick translation rate may vary!
The standard LMP protocol for SOLiD utilizes a fixed time for nick translation to control the size of LMP amplicons. Too short and your sequence reads will hit adaptor on their distal ends, too long and ePCR will fail to produce bright template beads. We use a 10 minute extension, instead of the 9 minute one in the protocol but even that fails to produce sufficiently long amplicons in some cases. What sorts of factors play into this variability? The obvious one it temperature. Even a little bit above 4 oC and the DNA polymerase will synthesize faster. Alternatively damaged DNA (eg, pyrimidine-pyrimidine dimers) might slow or stop the polymerase. We have recently seen results however that suggest that some other factor, more intrinsic to the genomic DNA itself may be playing a role.
Here are 7 LMP libraries we have attempted to construct over the past 9 months or so. I have given each library a number, "1-7". The following is a table with columns A-D. (Sorry, I did not see a better way to make a table.)
A Mean pre-circularization fragment size after hydroshear and gel cut (kb).
B Amount of post circle DNA used in nick translation (ng).
C Mean amplicon size after preparative PCR (bp).
D Mean "tag" size after subtracting out adaptors and dividing by 2 (bp).
All the libraries are from various rice species except for 2 and 3 which are chicken.
Libraries 6 and 7 are too amplicon small and are both from the same species and accession of rice. Note that the input DNA sizes and amounts varied as did the DNA prep, source of plant material and person that did the DNA prep. In addition the libraries were made roughly 6 months apart. Nevertheless the calculated tag lengths are nearly identical (50 bp) and are less than 1/2 the lengths of those made from chicken DNA.
Only a couple of data points here, but they suggest that there is something intrinsic to the genomic DNA that drastically alters the nick translation speed of the DNA polymerase being used.
We are making a new library this week...
Phillip
(1) Input DNA quantity and "quality".
I did a stint in an anaerobic protein purification lab early in my career. I was reliably informed by a post doc in this lab that even after a several thousand-fold purification of a protein of interest, to a single band on a stained gel, that every single protein present in the original cells was still in the purified prep. I'm sure this was a little hyperbolic, but the point still stands: purification protocols deplete a sample of the stuff you do not want--they do not completely remove it. Fortunately most of the hitch-hiking biomolecules in a sample DNA prep can be safely ignored. The ones that might trip you up:
RNA. Typically >90% of the nucleotides in a genomic "DNA" prep derive from RNA. (I saw 99.7% less than 1 month ago.) It is usually degraded down to oligos with RNAse so that you may not see it at all on a gel, but it confounds any attempt to measure DNA concentration using UV spectrophotometry. Be particularly suspicious if the sample reads > 1ug/ul concentration (genomic DNA is difficult to dissolve at that high of a concentration) or your 260/280 ratio goes above 1.8.
phenol. Less commonly used these days, but you will still see it. Just a tiny amount will throw your UV spec reading way off! 1 ml of water with 1 ul of phenol dissolved (0.1% v/v) into it will read like 540 ng/ul DNA. But if you look at the full spectrum it is obvious because the peak is at 270 nm.
nucleases. These are a little more problematic. Any self-respecting cell will be replete with all manner of nucleases. DNA preps generally should deplete them down to a tolerable level. But here is the rub: what constitutes "tolerable"? Add a little EDTA (as in TE) around to chelate divalent cations and most DNA nucleases will not function. However, bump the concentration of Mg++ high enough or remove the EDTA and they will be back in action. Since I am going to fragment my DNA using a hydroshear anyway, I don't mind the genomic DNA being a little smeary. But you really want to make sure all the nucleases are gone before you start ligating adaptors on.
(2) Your nick translation rate may vary!
The standard LMP protocol for SOLiD utilizes a fixed time for nick translation to control the size of LMP amplicons. Too short and your sequence reads will hit adaptor on their distal ends, too long and ePCR will fail to produce bright template beads. We use a 10 minute extension, instead of the 9 minute one in the protocol but even that fails to produce sufficiently long amplicons in some cases. What sorts of factors play into this variability? The obvious one it temperature. Even a little bit above 4 oC and the DNA polymerase will synthesize faster. Alternatively damaged DNA (eg, pyrimidine-pyrimidine dimers) might slow or stop the polymerase. We have recently seen results however that suggest that some other factor, more intrinsic to the genomic DNA itself may be playing a role.
Here are 7 LMP libraries we have attempted to construct over the past 9 months or so. I have given each library a number, "1-7". The following is a table with columns A-D. (Sorry, I did not see a better way to make a table.)
Code:
A B C D
1 3.5 516 350 120
2 5.1 500 327 109
3 4.6 500 326 108
4 2.5 446 300 95
5 4.7 240 250 70
6 3.0 177 212 51
7 4.6 530 209 50
B Amount of post circle DNA used in nick translation (ng).
C Mean amplicon size after preparative PCR (bp).
D Mean "tag" size after subtracting out adaptors and dividing by 2 (bp).
All the libraries are from various rice species except for 2 and 3 which are chicken.
Libraries 6 and 7 are too amplicon small and are both from the same species and accession of rice. Note that the input DNA sizes and amounts varied as did the DNA prep, source of plant material and person that did the DNA prep. In addition the libraries were made roughly 6 months apart. Nevertheless the calculated tag lengths are nearly identical (50 bp) and are less than 1/2 the lengths of those made from chicken DNA.
Only a couple of data points here, but they suggest that there is something intrinsic to the genomic DNA that drastically alters the nick translation speed of the DNA polymerase being used.
We are making a new library this week...
Phillip
Comment