Confusion over 'Not assayed' expression option

[jseager7]

At PHI-base we've been trialling the curation process with more publication authors, and at least one of them has been confused by the wording of the 'Not assayed' expression level:

My understanding is that in the publication, the presence of the gene product (expressed by the gene) was assayed, but not the level. Presumably, the author misinterpreted the 'Not assayed' label to mean that nothing was assayed, not even the presence or absence. In later discussion, the PHI-base team suggested that a more obvious label would have been 'Expression present, level not assayed'.

(I'm presuming the inverse case, 'Expression absent, level not assayed' is redundant because the level can't exist when expression is absent, and this is already covered by the 'Null' option.)

My suggestion would be to rename 'Not assayed' in this window to 'Level not assayed' or the full 'Expression present, level not assayed' (assuming the latter wording doesn't alter the intended meaning).

This should not affect allele and genotype names, where the short form 'Not assayed' will still be used.

@ValWood What are your thoughts on this? Have any of your community curators ever been confused by this wording?

@Hsinyugithub Please can you check that I've described the problem correctly?

Related: PHI-base/curation#228

[Hsinyugithub]

@jseager7 That's correct. @ValWood Also, the expression level box needs to be ticked to continue the curation process, so we want to make this as clear as possible.

[ValWood]

Hi, sorry for the delay, I'm on holidays.

We haven't had a big problem with this, but I suspect because all of our community curators are regular database users they understand the intention.

I wouldn't mind if it said explicitly
"product level not assayed" although it seems a little redundant.

Expression present, level not assayed would be incorrect.
Sometimes, a mutant may destabilise the protein so that it is immediately degraded. If the protein's expression level is not assayed, we do not know if the mutant is behaving like a knockdown. So this option is just saying "we did not asssay the expression level"

[jseager7]

I wouldn't mind if it said explicitly
"product level not assayed" although it seems a little redundant.

I agree about the redundancy, but 'product level' is already being used in 'Wild type product level', where presumably this could just be written as 'Wild type'.

My only suggestion would be to flip the wording to 'Not assayed product level' because I think changing the initial words of the label will be more distracting to users who are used to seeing 'Not assayed'. And also because 'product level' is at the end of 'Wild type product level'.

Here's the comparison of having the text before:

And after:

[ValWood]

I'm not sure. I agree, my suggestion is inconsistent.
However "not assayed product level" reads oddly

@PCarme / @kimrutherford do you have any suggestions

Another option is to change the "Expression" heading to "Expression level" and
remove the word "level" from "wild type"

[kimrutherford]

@kimrutherford do you have any suggestions

Sorry, I don't really understand the issue. Expression "Not assayed" seems clear to me?

In later discussion, the PHI-base team suggested that a more obvious label would have been 'Expression present, level not assayed'.

That sounds different to "Not assayed" because they established that expression is present? Which would need an assay?

[ValWood]

'Expression present, level not assayed'. would be incorrect.
(because expression may not be present, see my comment above)

Can I ask @Hsinyugithub @jseager7 how many curators have had a problem with this so far.

Is it a one off? or are people continually struggling with this?

I'm also wondering if the problem is somehow something specific to pathogen experiments in particular.

I'm also wondering if a more specific change might help. Like putting not assayed at the top, closer to the original "expression"
Sometimes people move down a list, and forget precisely what they are doing ;) Although to us the bottom seemed more obvious because you would only be able to use this if it is none of the other options.

[jseager7]

how many curators have had a problem with this so far.

I think it's just one so far, but I'll let Hsin-Yu confirm.

I don't have any objections to changing the Expression heading to 'Expression level', but the help popup already makes it clear the section is referring to expression level:

Like putting not assayed at the top, closer to the original "expression"

I'd prefer not to move the option somewhere else in the list, because this is likely to interfere with muscle memory for experienced curators.

I suppose the question here is, if a curator gets confused or stuck, are they going to choose 'Not assayed' as the closest option, or are they going to not make the allele at all? If they fall back on 'Not assayed' then there's no problem. Perhaps we shouldn't rush to make a change unless it's actually stopping community curators from curating.

[Hsinyugithub]

Sorry for the late reply—I was off on Thursday and Friday. The issue isn’t about where to find the "not assayed" option; it's more about the type of curation we do in PHI-base. We have many effectors that are extracted from the pathogen or an expression system and then infiltrated into the host. So, it's difficult to describe this in terms of expression level per se. The authors might report the concentration of the infiltrated peptides, but I wouldn't describe that as expression level. In such cases, we would select "not assayed." However, it feels a bit odd to use "not assayed" in an experiment where we know the peptide is present, but it just can't be described in the conventional way.

I think "Expression present, level not assayed" or "Expression level not assayed" would be a bit clearer. I’ll discuss this with Kim and Martin to see what they think.