Marina, if it is a pure species eluting at that 'dimer' position then you should observe a plateau in the molar mass chromatogram there. I suspect that is not the case. Remember, MALS measures the weight-average molar mass of whatever is in the detector at that moment. If you have a little bit of a VERY high mass species co-eluting with a lot of dimer, the weight-average value can be much higher than the dimer mass. So in short there is nothing wrong with the MALS detector, it is a chromatography issue.

Often very high mass aggregates are quite sticky and will elute in a long tail extending across nearly the entire chromatogram. When the influence of such species is not too large this will produce molar mass chromatograms that "smile" across each peak (the mass is higher in the wings of the peak).

But it is also possible that the very high mass species that are influencing your chromatograms aren't really present in the sample you are injecting, but are things being knocked off the column by the injection itself (by the pressure pulse, or perhaps by components of the formulation buffer for the sample you are injecting (perhaps detergents). Definitely you want to do a buffer blank injection to check whether that is producing significant scattering signals.

I also suspect you are trying to measure minor peaks at very low levels. The reality is that the aggregate levels in many monoclonal antibody products are just too low to get good MALS data for the minor peaks. If you want to identify the species you may need to use stressed samples to push up the levels of the minor components.

Thank you very much, John. What I actually want is to identify these species. If these are protein aggregates (3000 kDa), why are they eluting together with the dimer? Or is it something else? In any case, it is a good idea to start with a buffer injection and take things from there. Thank you again.

Marina, if SEC elution position was a reliable way to determine protein masses then there would be no need for the MALS detectors! Protein species often bind weakly to the SEC resin and therefore elute much later than expected based on their true mass.

Aggregates may expose hydrophobic patches which normally are buried in the protein core and can be much stickier than the monomer.

You haven't described the molar mass profile across this peak eluting near the expected dimer position, and that is a key diagnostic for what is going on. Do you see a plateau at 3 MDa across the center of that peak, or does it 'smile' up on either side, or do you see a continuous drop across that region?

As I said, very large protein aggregates often elute as a peak that starts at the void volume, but then falls off with a very long tail, which may actually extend through the entire chromatogram. Indeed sometimes you don't see the MALS signal return to baseline until long after the salt peaks elute (perhaps not for 2-3 column volumes). When you have this phenomenon, this tail of very high mass material pollutes other minor peaks and can substantially influence the mass chromatograms.

Hi John, No, there are no smileys in those data. The MALS signal goes up right after the monomer peak (to 20 times higher MW), then makes some kind of a flattened camel hump shape and goes back down again.

You are right about the low levels in the aggregate species - these are about 1 %. Also there is a tail throughout the chromatogram, as you mentioned. So I tend to think that the high MW values are due to the pollution as you stated in the last paragraph. Thank you again, John.

Potential reasons for this discrepancy could be:

a) the dn/dc is actually different for the dimer of the two variants

b) the dimer concentration is so low that the effective error of the molecular weight is larger

c) an aggregation equilibrium where the dimer elutes and then reforms higher order oligomers again

To investigate, see if running at twice the injection concentraiton shows the same elution profiles again. Also try batch DLS, if the variants are the same, then the DLS batch result should also be the same, if they are different, the DLS for the second sample will also support the SEC result.

Thank you, Ulf, I'll try that.