I promise not to be short...

I tried different columns: Eksigent, Agilent, Thermo, self-packed columns with Agilent 1.7 um C18 sorbent and Phenomenex ordinary C18 3 um sorbent (no core-shell or any such stuff - just did not have a chance yet). I also tried NewObjective ready-made PicoFrit columns. In my experience I did not see any really important differences between the column effectiveness. As Varatharajan Sabareesh stated it really depends on your task, whether you want to identify, to quantify a few or to quantify everything. I usually use max 120 min gradients for 15 cm*75um columns with about 1500-2000 proteins in the full Ecoli digest identified with 1%globalFDR (with TripleTOF5600 system and ProteinPilot software). I also tried 25 cm PicoFrit columns with 4 h gradient and was able to identify about 3000 proteins, though I did not make any real optimization for this column yet. The problem with 25 cm PicoFrit column with NewObjective own sorbent ProteoPepIII was an extremely high backpressure (actually baturally expected, around 4500-5000 psi) even at lower flowrate (I ususally use 300 nl/min for 15 cm columns with about 100-150 bar, with this column I had to stick at 200 nl/min to be able to work at all at 4500 psi) and I could do nothing about it since NanoSourceIII (ABSciex) does not support in-source column heating. It was also not that easy to mount a long emitter-packed column on this source (and a new version of that source with the full plastic cover would make such columns just impossible to use, a pity for future customers). The problem of such column is that it is an emitter in itself, so when the tip goes wrong - it is dead, and the tip is much easily damaged than the column itself. i had to increase the voltage for NanoSourceIII from 2500 V to 2700 V to get the same sensitivity level as with a normal emitter. To my disappointment I did not see peak narrowing (I hoped for having no post-column dead-volume) but I think that was because I had to decrease the flow-rate.

I do dislike Agilent columns because of their large apparent post-column dead-volume - they use unions to pack a column between them with a usual metal frits at the end of a column. It is good if you have a frit blocked - you can change it or wash it, but it is just that I do not like it that way. Anyway when you have a normal column blocked you can cut a small piece of it with some sorbent from the front end of the column - it works perfectly with Dionex columns (and they do really like to get blocked or partly-blocked). It is however impossible with Agilent columns that are packed within not fused silica, but within PEEK-silica.

With the flow rate of 300 nl/min all the columns tried gave the peak width at half-height about 8-15 sec and the full peak width at the peak base around 20-30 sec (depending on a lot of factors). However in my biased opinion Agilent columns gave it slightly wider... With the best results (for my system) with Eksigent columns due to some mounting things (below). And the most excellent results one can get with chip-systems (either Agilent Chip-cube system, or ABSciex cHIPLC - these chip systems differ in the way what the chips contain - in case of Agilent system they are a full trap-chip-metal emitter pack, in case of ABSciex the chip contains either a precolumn or a column). The plus of an Agilent chip is the simplicity of its use - so when you have an unlimited money supply you can stick to it. But if any part of a chip is blocked you have to throw it away (very expensive). For ABSciex chips - they are just made to make all the dead volumes lower and they cost about the same price as their columns.

When you choose a column you should take into account the following factors:

1. Your task and, correspondingly, the gradient length you would like to use:

if you want great identification - long columns - long gradients - low RT reproducibility  and wide peaks at the end of a gradient - but who cares

if you want MRM quantitation - short gradient - short columns - good RT reproducibility. You can even think of using shorter than usual 15 cm columns, you can make them youself by cutting a commercial column by yourself - faster reequilibration

2. Ease of mounting for your system:

I had a nightmare trying to attach 290 um OD Dionex columns to my system that was designed for 360 nm. And the sleeves for Dionex capillaries (just because the manufacturers like Idex make just several defined IDs with a step of more than ) end up having pretty large free space between the capillary and the sleeve, that is not good.

The same thing goes with Agilent columns that are supposed to be attached via special Agilent Nut-ferrules that have 1/32 ID and 1/16 OD. There is a lot of opportunities for making post-column dead-volumes. It is also funny that turned out impossible to pressure-tight these ferrules with any capillaries (even PEEK-silica) but the Agilent ones - slight differences in the material or OD may be.

It is also necessary to think whether you would be able to thermostat the column you want with the hardware you have. For 50 cm columns to be heated you would have a very special (or spacious) thermostat and you would not be able to go with such column without it because of back-pressure.

3. Column and precolumn compatibility

You should take into account that it is the best practice to have the same sorbent in a column and in a precolumn. If you take a PicoFrit emiiter column with a cool core-shell sorbent, but you have a usual C18 PepMap sorbent in a precolumn it is possible you would lose all the advantages due to slight differences in sorbent surface peptide binding.

The question of sorbent choosing is a very interesting one. Besides, it is very interesting to see some real comparison between defferent sorbents in application to LC-MS/MS experiments and not just their great peak capacities and resolution and low back-pressure and so on, that the manufacturers promise. But the funny thing is that in the reports and bucklets (and manuals) of some new sorbent the manufacturers compare it with some "ordinary sorbent of other manufacturer" - just show me an ordinary one - nobody does it in ordinary way.

I use Acclaim™ 3μm PepMap™ 100 Nano LC Columns, 75um I.D. since ten years (despite I have tried a lot of different columns) because I can get a good separation with good reproducibility and robustness, of the nanoHPLC system (Ultimate 3000, with splitter 1:1000). But if I had a UPLC system, I would change the columns.

Thanks for the input Francesco! Have you ever tried solid core particles? They are the hype these days in the conventional size LC world, but I haven't seen much use with 24/7 proteomics labs.

Thank you Steven for the interesting question. I tried them and use them for complex multiprotein samples (at high concentration). I do not like the lack of back pressure and sensitivity when work on individual spots from two-dimensional gel that are the main task in my lab.

...I think that proteomics columns will soon be 10-20 um inner diameter. This will give better spray and less suppression, and lower radial dilution. But will the manufacturers soon make pumps that can easily handle 5-20 nl/minute?

It depends what you want to separate. C-18 will have its limitation but for routine work silica monolith would be OK (but not the best).

The choice of column depends on the type of question or research project that one focuses on. There are two important aspects in proteomics: 1. to identify many proteins, i.e., maximize identifications and 2. quantitative and comparative proteomics. In the case of maximizing identifications, depletion of well-known top abundant proteins needs to be carried out, for which immunodepletion columns may have to be used. Alternatively, some precipitation methods may also be followed, so that the experiments can be focused on low abundant proteins and thereby the identifications can be maximized. To work with low concentration or low volume samples, capillary LC or nano LC (reverse phase) may be suitable, which have better sensitivity than normal/conventional LC. Normal/conventional LC (reverse phase) means it would involve columns of dimensions, 2.1 mm x 100 mm or 4.6 mm x 100 mm. Further, in the case of bottom-up approach, which mostly deals with proteolytic peptides of molecular mass not exceeding 2 or 3 kDa, C18 columns would suffice. Else, if one is going to work with peptides of molecular mass more than 2 or 3 kDa and up to 6 or 7 kDa, i.e., middle-down approach, then C8 columns can be tried. Conventional LC involving column of dimension 2.1 mm x 100 mm or 4.6 mm x 100 mm, may be suitable for quantitative or comparative proteomic experiments, which would not involve low concentration or low volume samples. With regard to choice of column material, monoliths with higher porosity would give less back pressure compared to standard packed columns of particle size 3 micron or sub 2 micron particles. Prior to carrying out reverse phase LC, it may be good to perform ion exchange or immobilized metal ion affinity chromatography (IMAC), which can aid in maximizing identifications.

I use Kinetex core-shell technology columns. They work nicely for my samples.

Steven, in my experience it all depends on context. What flow rates are your HPLC and ion source comfortable with? Can you self-pack or do you have to stick with commmercial columns? If so, how price-sensitive are you? What kind of samples do you typically analyse (2D gel spots or whole cell lysates), and into which gradient length does this translate? How fast is your MS instrument in acquiring data?

In our lab we operate thrre to four different kinds of columns, based on the answers to the above questions.

Hi Christof! In our current study of columns, we are operating in the 100-200 nl/min flow range. Usually our samples are full cell lysates, and doing targeted analysis, with a time limit of 30 minutes. We currently use a Q-Exactive (but will soon have a qqq).

What would you use in this context?

Would e.g. peak capacity be a big issue for you, or do you find that this is not as big of a deal as many chromatography people would insist?

+ has anyone here tried HILIC columns for comprehensive proteomics? Check out monolithic HILIC columns here:

http://www.ncbi.nlm.nih.gov/pubmed/?term=tanaka+hilic+proteomics

If you do not want to perform gradients much longer than 30 min then there are a couple of nice options especially for targeted work. I have had good experience (peak width 5-6 s FWHH) with Waters Acquity TSS-3 25 cm (expensive though! and hgigh back pressure, so need a column oven) and Merck Chromolith SpeedRod C18 (low back pressure), however at higher flow rate. I like to run 75um columns at 300 nl/min to reduce effects of diffusion and dead volume. For cost reasons though we would usually self-pack 20cm with Dr Maisch C18aq 3um - for short gradients, performance is not far behind the theoretically fancier commercial columns.

Thanks Cristof for your detailed answer! Really appreciate the input.

We are using  a  C18 PepMap, 3 micr, 100 A,  15 cm x 75 micr  at a flow of 300 nl/min  with very good results.

Best regards

We pack our own columns. For flow rates of 300 nL/min, I use a Picofrit column from New Objective (the tip is fritted and incorporated into the column rather than separate and joined with a union), 75 µm I.D. with a 15 µm tip. We use Dr. Maisch ReproSil-Pur C18 AQ, 3µm diameter with 120A pores. I usually pack my columns to 15 or 20 cm in length. Back pressure with 98% aqueous phase at 300 nL/min is usually ~120-150 bar.

Thanks Kristian! How many runs do you usually use these home-packed columns for? What usually makes you replace the column (e.g. bad spray, bad chromatography, high pressure)?

I don't really have a solid upper limit for the number of runs. Since columns packed in-house are relatively inexpensive, I usually pack a new column for each new batch of experiments. I have used a single column for many days in a row, even up to two weeks of continuous acquisition on large data sets.

The column life time is usually limited by the tip. If spray gets poor, it can often be fixed by simply carefully wiping the tip with a Kimwipe and methanol, and/or by increasing the spray voltage. I replace the column if the signal becomes unstable, or if signal is unacceptably low as determined by a glufib standard run interspersed between samples. I have noticed that sometimes the spray doesn't look perfect, but that doesn't necessarily mean that ionization is poor. I always use signal strength and signal stability to determine if my tip needs to be cleaned or replaced.

I also replace the column if the back pressure gets too high, but this is usually due to some grit from a dirty sample or a deteriorating rotor valve, not the column itself degrading. If the tip is still good, the column can be back-flushed and re-packed.

I rarely see the chromatography degrading in the lifetime of my column if my samples are clean. I rarely see peak tailing due to dead volume as I often see arise when I have a tip attached to the column by a union. Even if the tip is securely fitted at first, over time the sleeve can fail. I've even seen the polyimide delaminate from the fused silica, creating a dead volume.

It may be more expensive to pack new columns as often as I do (as opposed to using the same column for a long time and swapping tips), but I find that I waste less time troubleshooting and lose less precious sample this way, so to me it's worth it.

We are using a C18 PepMap, 2 um, 100 A, 50 cm x 75 um at a flow of 200 nl/min. I advice you for the use of long columns for complex proteomes.

Hi Kristian, thank you for your post! Really good point regarding the advantages on no needing a union between the column and tip. We are currently working with 10 um ID columns and the dead volume here can be a challenge. Will check out if New Objective will have any Picofrit plans for these dimensions.

Enrique uses 50 cm columns (thanks for the post Enrique!!). Kristian, have you ever used so long columns self packed?

Enrique, how high pressure do you 50 cm columns operate at? Have you ever tried/thought of trying long monolith columns (lower back pressure)? Or do you feel this isn't really a big problem?

we synthesize the C-18 bounded sub 2 um silica stationary phase and pack our own   columns of various dimensions in our lab, efficiency and resolution of our column are much more better than commercially available C-18 column as we have reported in several papers in journal of chromatography (A) 

I have desalted peptides in a trap column (180 μm × 20 mm, packed with C18 Symmetry, 5μm, 100Å (Waters, Taunton, MA)) and subsequently resolved them in a reversed phase column (75μm x 250 mm nano column, packed with C18 BEH130, 1.7μm, 130Å (Waters, Taunton, MA)) using a 4 hour gradient.  This is done on a Waters NanoAcquity interfaced to a Thermo Orbi (XL, Velos, or Elite).  I usually inject about 400ng of peptide digest.

Thank you Ali for your input! Will definitely check out your work.

Dear Sara, appreciate your comment, will check out this longer UPLC column!

I promise not to be short...

I tried different columns: Eksigent, Agilent, Thermo, self-packed columns with Agilent 1.7 um C18 sorbent and Phenomenex ordinary C18 3 um sorbent (no core-shell or any such stuff - just did not have a chance yet). I also tried NewObjective ready-made PicoFrit columns. In my experience I did not see any really important differences between the column effectiveness. As Varatharajan Sabareesh stated it really depends on your task, whether you want to identify, to quantify a few or to quantify everything. I usually use max 120 min gradients for 15 cm*75um columns with about 1500-2000 proteins in the full Ecoli digest identified with 1%globalFDR (with TripleTOF5600 system and ProteinPilot software). I also tried 25 cm PicoFrit columns with 4 h gradient and was able to identify about 3000 proteins, though I did not make any real optimization for this column yet. The problem with 25 cm PicoFrit column with NewObjective own sorbent ProteoPepIII was an extremely high backpressure (actually baturally expected, around 4500-5000 psi) even at lower flowrate (I ususally use 300 nl/min for 15 cm columns with about 100-150 bar, with this column I had to stick at 200 nl/min to be able to work at all at 4500 psi) and I could do nothing about it since NanoSourceIII (ABSciex) does not support in-source column heating. It was also not that easy to mount a long emitter-packed column on this source (and a new version of that source with the full plastic cover would make such columns just impossible to use, a pity for future customers). The problem of such column is that it is an emitter in itself, so when the tip goes wrong - it is dead, and the tip is much easily damaged than the column itself. i had to increase the voltage for NanoSourceIII from 2500 V to 2700 V to get the same sensitivity level as with a normal emitter. To my disappointment I did not see peak narrowing (I hoped for having no post-column dead-volume) but I think that was because I had to decrease the flow-rate.

I do dislike Agilent columns because of their large apparent post-column dead-volume - they use unions to pack a column between them with a usual metal frits at the end of a column. It is good if you have a frit blocked - you can change it or wash it, but it is just that I do not like it that way. Anyway when you have a normal column blocked you can cut a small piece of it with some sorbent from the front end of the column - it works perfectly with Dionex columns (and they do really like to get blocked or partly-blocked). It is however impossible with Agilent columns that are packed within not fused silica, but within PEEK-silica.

With the flow rate of 300 nl/min all the columns tried gave the peak width at half-height about 8-15 sec and the full peak width at the peak base around 20-30 sec (depending on a lot of factors). However in my biased opinion Agilent columns gave it slightly wider... With the best results (for my system) with Eksigent columns due to some mounting things (below). And the most excellent results one can get with chip-systems (either Agilent Chip-cube system, or ABSciex cHIPLC - these chip systems differ in the way what the chips contain - in case of Agilent system they are a full trap-chip-metal emitter pack, in case of ABSciex the chip contains either a precolumn or a column). The plus of an Agilent chip is the simplicity of its use - so when you have an unlimited money supply you can stick to it. But if any part of a chip is blocked you have to throw it away (very expensive). For ABSciex chips - they are just made to make all the dead volumes lower and they cost about the same price as their columns.

When you choose a column you should take into account the following factors:

1. Your task and, correspondingly, the gradient length you would like to use:

if you want great identification - long columns - long gradients - low RT reproducibility  and wide peaks at the end of a gradient - but who cares

if you want MRM quantitation - short gradient - short columns - good RT reproducibility. You can even think of using shorter than usual 15 cm columns, you can make them youself by cutting a commercial column by yourself - faster reequilibration

2. Ease of mounting for your system:

I had a nightmare trying to attach 290 um OD Dionex columns to my system that was designed for 360 nm. And the sleeves for Dionex capillaries (just because the manufacturers like Idex make just several defined IDs with a step of more than ) end up having pretty large free space between the capillary and the sleeve, that is not good.

The same thing goes with Agilent columns that are supposed to be attached via special Agilent Nut-ferrules that have 1/32 ID and 1/16 OD. There is a lot of opportunities for making post-column dead-volumes. It is also funny that turned out impossible to pressure-tight these ferrules with any capillaries (even PEEK-silica) but the Agilent ones - slight differences in the material or OD may be.

It is also necessary to think whether you would be able to thermostat the column you want with the hardware you have. For 50 cm columns to be heated you would have a very special (or spacious) thermostat and you would not be able to go with such column without it because of back-pressure.

3. Column and precolumn compatibility

You should take into account that it is the best practice to have the same sorbent in a column and in a precolumn. If you take a PicoFrit emiiter column with a cool core-shell sorbent, but you have a usual C18 PepMap sorbent in a precolumn it is possible you would lose all the advantages due to slight differences in sorbent surface peptide binding.

The question of sorbent choosing is a very interesting one. Besides, it is very interesting to see some real comparison between defferent sorbents in application to LC-MS/MS experiments and not just their great peak capacities and resolution and low back-pressure and so on, that the manufacturers promise. But the funny thing is that in the reports and bucklets (and manuals) of some new sorbent the manufacturers compare it with some "ordinary sorbent of other manufacturer" - just show me an ordinary one - nobody does it in ordinary way.

Dear Sergey, thank you for your excellent and well guiding input! I share many of the same experiences as you describe here. Great that you discuss the choice of SPE vs. LC material here. In our experience, it can often be a good idea to use an SPE with a slightly lower retention factor than the LC column, to ensure refocusing when the analytes are eluted on to the LC column.

Dear Steven, what kind of sorbents (and precolumns) do you use for trapping material with lower RT factor? I do have C8 Agilent precolumns I tried to use with my C18 columns (of different manufacturers) but actually I somewhat dislike Agilent for stupid connections (1/16OD ferrules). Besides, these precolumns are pretty big (5*0.3 mm) that certainly gives good capacity, but it also gives a lot of dead time before the gradient finally comes to the sep. column.

By the way, there is also an interesting thing with precolumns I noticed: if one uses a small precolumn (e.g. 0.5*0.3 mm), it does not matter what would be the flow direction for gradient pump from this precolumn and it is much better to make it the same way as a loading pump has loaded material to the precolumn (it can help to avoid column blocking by some material that went through the filter between the loop and the trap column. But when one uses a big precolumn, it is better for one's resolution to configure the valve the way the gradient pump runs in the opposite direction from the loading pump. This way the analytes do not go through the whole precolumn volume if they do not have to and that helps a little with peak width. However it is better to install an additional filter before the anal. column. 

Hi Sergey! For SPE and LC, we often use Kromasil C8 and Kromasil C18, respectively.

These are packed for us by GT Septech (Kolbotn, Norway). They pack the SPEs in "Tracy" columns, which are available in e.g. 0.3 mm (ID) X 0.5 mm (length). These guys are really good!

I agree that frontflushing and backflushing are both good choices, especially for small SPEs.

For filtration before SPE, we have designed a plumbing scheme, that ensure that the filter never clogs up, allowing for hundreds/thousands of runs. Check out e.g.

http://www.jlr.org/content/early/2014/05/02/jlr.D048801.abstract

Hi, Steven. It again shows that whatever you think nobody thinks of and cares for just in reality is at work in some place (just because there are so many people that it is really hard to make up something really original :)) ). And I really thought of trying to manufacture some sort of a short letter concerning the idea of C8 precolumn - C18 anaytical column pairing - just because I do not look for such articles in lipid studies, working with peptides. However for some reason the manufacturers do not usually do lower RT factor precolumns (e.g. C8)... But, no, they do make them, but it is not widely used for C18 analytical LC. I wonder why.

By the way, if your aim in the article was to get higher sensitivity  for your lipids why did you use such large ID analytical column? If you take it smaller, usual 0.075 or even better 0.05 mm instead of your 0.15 mm you would greatly gain in sensitivity. Though you would have to use about 150-200 nl/min for 0.05 columns instead of 500 nl/min and it would be better to use smaller precolumns - you would wait for ages for your analytes to get to the separation column with a 0.3*5 mm precolumn.  

Hi Sergey! I definitely think there would be room and interest for a study showing improved refocusing of peptides with C8-C18. Why manufacturers do not take advantage of this trick is rather surprising.

Where did it say that the column was 0.15 mm? It was 0.1 mm ID 15 cm long. But you are right, smaller diameters would be better. But in the lipidomics world, the columns we use are already extremely narrow (more usual with 1-2.1 mm ID). It is so funny how the proteomics community are much more comfortable with narrow columns than the drug/metabolomics crowd. This is a cultural difference-I think that all fields would benefit from very narrow columns. PS. On our plan is to employ open tubular columns for metabolomics, with 10 um IDs. We have had success with this in proteomics, and will soon have a paper published on this.

Hi, Steven!

Do you mean the columns without a frit where the sorbent  is stucked by itself due to the particle size comparable to the tubing ID (there was a ratio for it somewhere)?

10 um would make a bomb... if you can get the flowrate low enough and the spray stable enough...and the best of all would be to get some sub-um sorbent...can start with grinding 1.8 um particles to dust (making them "duster") :))), otherwise there would be very small amount of the sorbent in the column and the capacity would be disastrous, only if you do not make it 100 cm length... Well, I would be waiting for the article.

150 um - my err - while quickly getting through the methods confused 150 mm length with 150 um. And afterwards sometimes wondered how it is possible: 150 um and only 500 nl/min. Now its clear...:)

Sergey, here is a preprint of our latest 10 um columns efforts:

Article Open Tubular Lab-On-Column/Mass Spectrometry for Targeted Pr...

Hi, Steven! Thank's a lot. 

Hello, Steven. Went through the article... Well, first of all - not what I expected :) - I mean you used absolutely different column design with 10 um ID than the ordinary one, however... In advance sorry for being verbose - I tried to explain my ideas as thoroughly and precisely as I could.

The article is interesting. It really goes farther than the previous ones in identification efficiency with on-line trypsinolysis. And it is really great to have an option of on-line digestion coupled with the analysis. By the way, you did not emphasize it but I think that the greatest value for such setup is that you can standardize trypsinolysis for very low protein amount (actually, whatever low protein amount you want) that is challenging for common trypsinolysis that does not go very well with ultra-low total protein concentration.

However I have several remarks. When I completed reading I had a huge pile of questions:) that were not dealt with. The article feels like you had two goals: to optimize sensitivity through very narrow column ID and to optimize trypsinolysis for on-line use. Unfortunately, I cannot claim to be a great specialist in PLOT columns so I cannot say how original its use for proteomics is, but I feel like wanting more of it in the article (just more statistical data about its efficiency, peak capacity and so on). So:

1) Trypsinolysis

it is very interesting to know the real efficiency of this trypsinolysis setup. Even without the sophisticated column setup used in the article. Something like a comparison between "OTER trypsinolysis+normal analysis" and "common trypsinolysis method (e.g. RapiGest)+normal analysis". Actually my main question is how efficient the OTER  trypsinolysis is. It is very well that you identified 1500 proteins, but identification is not that important in comparison with "correct" quantification (at least for peptides). There are a lot of works on trypsinolysis efficiency and the time and conditions adustments. It is well known that all the trypsinolysis methods give somewhat different peptide ratios (and peptide lists) but it is always accepted that you need to make it as completed as possible - with no uncleaved protein parts left behind. Usually it is dealt with a dependance between peptide concentration and e.g. time. When the peptide concentration stops to increase it is considered that for this peptide the trypsinolysis goes up to the end (this time is always different for different peptides). In my opinion it is a little bit strange that you take 5 ul of the sample and put it through 1.2 ul reactor at a flow rate of 0.5 ul/min. Unfortunately, I do not know (yet) what is the principle of OTER trypsinolysis (at least what it is believed to be), but for me it is obviously unclear:)). In off-line trypsinolysis with the use of some support for trypsin (either bead or a column setup) the support works as a support for trypsin only (as far as I know, believe and understand) - and it really matters what is the time of protein incubation (with the protein in the solution) with such a trypsinilated support. But here it is obvious that you expected the OTER column to hold the proteins (and may be the generated peptides as well) since you put in waste most of the original sample volume when it went through the OTER column, at a pretty high flow rate by the way, but are you sure it really did hold everything? And how much of protein material went to waste unaffected? And if you really suggested solid-phase reactions between absorbed trypsin and absorbed protein how can you be sure it really went well for all the proteins. Wasn't there any OTER column selectivity for proteins? Do you see the same proteins in the identification lists (with the allowance for pseudo-random IDA principle) as in case of usual trypsinolysis methods - that is do you really see the proteins according to their concentration in the original sample. Reading the article I expected all the time that e.g. different length OTER columns tripsynolysis efficiency would be compaired, different flow rates etc. - compaired quantitatively in terms of peptide EIC area. There is a link to Supplementary figures, however I was not able to download them through the ResearchGate link you provided. 

In summary, the main question is the proportionality of the peptide signal to the amount of protein loaded to the OTER column and the dynamic range of such proportionality. It is best to have several proteins protein mixed in different known concentrations with some complex sample, followed by OTER digestion. And to have a complex sample loaded in different amount (10-100-500 ng) and comparied in terms of EIC areas. 

2) HPLC setup questions:

precolumn peptide capacity

column peak capacity 

It is unclear what is the real setup advantage in comparison with the common column setup. If I take a cutting-edge tripleQuad system I can reach as low as several attomoles peptide quantification so that 10 ng of protein tells nothing without a direct comparison. Is it possible to analyze the similar sample with the use of an ordinary column setup (precolumn - 0.075*150 mm column) with the same MS system and to compare it with the PLOT one. In terms of s/n, peak width and anything else I forgot about? Just some reference point. 

Something like this:)))

Dear Sergey, I am terribly sorry for the belated response to your excellent comments!!

I have thought a lot about your points. The OTER column will now be thoroughly investigated for comprehensive proteomics (the main intent was targeted, which I am far more comfortable with), as well as getting the quantitative aspect in place (which method do you prefer?). 

Regarding advantages, we have seen that OTER/PLOT columns are simple to reproduce compared to small ID monoliths (becomes increasingly difficult to go down in ID with packed columns), provide high sensitivity due to low radial dilution (a comparison paper is in preparation), and give good peak capacities compared to commercial columns in small dimensions (Rogeberg et al 2013). Suprisingly, the capacity of the SPE-PLOT system was comparable with larger systems (Robeberg et al).

But the PLOT can/will get better when nanotips can be used in-column (not with unions, such a source of band broadening). The OTER will now be tested with regards to your thoughts on evaluating protein amount/peptide signal.

Really appreciate your thoughts, this was a more detailed critique than the peer reviews we got!

Hello all 

I am using C18 50cmx75um column in Nanolc 1200 at a flow rate of 300 nl/min, but i am found problem related to intensity & retention time can you suggest how come I solve the issue

Hi,

I`m wondering if I may re-open this discussion with a brief follow-up..

Sergey Kovalchuk mentioned above that longer columns have less RT reproducibility and broader peaks towards the end. Could anyone please explain why and would you have references?

Thank you!