Dear Flavio, Fatima, Gaia, 

congratulations for these analysis and for the impressive results. 
As Nikhef group we have discussed your paper.

We find your analysis very solid. We are however a bit concerned by the comprehensibility of your paper. Being the first time we present this analysis, some of the points would deserve a bit more space: there is no need to go for a longer paper as some other parts can be reduced.  
It is clear that all of the parts in common with the Bsmumu analysis can be shortened a bit (BDT(s) for example?) while we would suggest to concentrate on the differences. 

The paper would profit of the observed invariant mass distribution (or of the events in bins of BDT given this is a single bin search in the mass), maybe with a fake signal superimposed,  while the CLs plots could be additional material. 

Furthermore the paragraph to translate the BR into a mass limit would deserve some more  explanation (which you partially include in the motivations to PRL). It is in particular important to say that the two limits are given in two disjoint hypothesis (tau associated to the first or second generation) so that it should be clear that the two limits are for two different leptoquarks (otherwise only the best limit would be important). Maybe it should also be mentioned that a diagonal four dimensional CKM matrix is assumed. 

Below we write some more detailed comments on the paper. 
Some of them might be superfluous depending on how you implement our general comments. 

Again congratulations, 
Regards, 
 Francesco, on behalf of the Nikhef and VU group.


Detailed comments: 
Abstract: here and in the conclusions the notation MLQ(Bs->emu) etc needs to be changed. The mass is not a function of the BR, the limit on it is. Here you probably mean the given hypotheses: see general comment.
Line 167: Remove "allow to".  It is either "allow to probe" or "provide a probe of".
175: the supersymmetry models -> sumpersymmetric models (remove the)
194: the luminosity (2 fb^-1) is not needed and can be removed since the energy and collision type are not given.
204-207: This comment could be moved later.
At this point we think it would be important to introduce the analysis strategy, introducing the concept of normalisation and control channels. Else the rest is hard to read.
244: B is undefined.
245-247: no need to be so vague. We suggest to say that bremstrahlung is 
partially recovered by using photon depositions in the ECAL.
246: bremsstrahlung (not capital) 
256-259: This sentence in general is not true. You could say that you require the Bdkpi to be selected by the same lines etc..
257: Add "event" in "This channel has an event topology similar ...".
267: Remove "the direction" in "the direction of the momentum".
270: - Write "the B candidate IP, its decay vertex chi2, ...",
        if that is what we mean here.  The original text reads as "the IP of the B candidate,  the chi2 of the primary vertex, ...".
- line 270: what is the difference between the chi2 of the B vertex 
(line 270) and the chi2 of the SV (line 272)?
301: Probably here you mean the invariant mass of the background is independent of the BDT. 
308: the B->hh channels have not been introduced before for what concerns the selection etc. 
330: It depends on event multiplicity and you use nSPD as a measure of 
it. But the signal shape does not depend on nSPD. Replace by event 
multiplicity everywhere.
358-363: this parameters are not needed (in particular the mass central values do not have to be in the paper) while they could be replaced by a mass plot
368: fluctuating with toy experiments -> toy experiments fluctuating
381: it is not clear to the reader here why do you need f_u ? It only becomes clear later, write it there.
382: the measured PT dependence is not negligible in itself. Please rephrase.
399: Use trigger paper instead of [30].
407: This sentence repeats the one in 282.
418: Probably here you mean the statistical uncertainty on the background not the systematic one. 
445: Here you should also mention the misidentification to electrons 
451-453: Move that to 442.
464: Remove "in the full bdt range" as it is repeated later
469-471: repeats 456
473: remove "carefully"
Table I: The line on "exp. excl. bkg (from MC) is confusing. The reader is tempted to sum up lines 1 and 2 to be compared with the observed in 4. A comment in the caption at least is needed or a more clear formatting of the table itself.
511: see general comment on the physics explanation 
513: some of these parameter values (certainly the B masses) can be removed leaving only the references 
524: also here the Z and top masses are not needed if references are given 
Ref 21: G. Lafranchi , A. Sarti and others --> G. Laranchi et. al.
   is this really the only reference for LHCb muon ID?