This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Forum: Guidelines / Limits for Shared Bioinformatics Support

I have a longer description of my thoughts in this collection of blog posts here, but I think these are the main questions where I believe I could benefit from getting some feedback:

1) Are you aware of any suggested limits for shared support (in terms of the number of projects / PIs supported by an individual)? For example, a colleague passed along this paper describing that Biostatics support should not go below 5% effort.

2) Are there any guidelines for how much time should be expected (per week or per month) to receive adequate guidance and carefully review results before publication?

3) I think there is some need to have labs knowingly collaborate with each other. I can understand that this will be more difficult as you support more labs. However, are there any guidelines about communicating concurrent projects and/or developing code that is the result of support for labs that may not be in communication with each other?**

4) My understanding is that key personnel are supposed to be listed on IRB and/or IACUC protocols (for example, to keep track of everybody who had access to a patient’s genomic data, understanding the requirements for depositing that public data for publication, avoiding unnecessary risk and/or pain in subjects, etc.), and amendments are required if there are changes from the original plan. Are there cores that have staff members (such as “Bioinformatics Specialists,” like myself) listed on grants and/or IRBs?**

Any help that I can get to understand other people’s experiences and communicate myself more effectively will be greatly appreciated!

core shared-resources collaboration
• 2,540 views • link
updated by Ram Ram • written by Charles Warden Charles Warden

1 answer

This is a complex topic and a lot depends on the local environment.

1- I am not aware of limits. I think common sense should prevail. Depending on the nature of the projects, it's possible to interleave them but sometimes, one just has to say no to new projects or give an estimate of a possible start date given current capacity.

2- In my work, I distinguish 4 types of interactions. If that helps, I can give the median effort required for each type but there are large variations between projects:

  • Consulting: most of the time this means one initial meeting and a follow-up meeting for a total of ~2 h
  • Data analysis, no need for extensive data exploration or new methods: 0.35 PM (person month)
  • Data analysis, requires extensive data exploration and/or new methods: 1.4 PM
  • Only IT work/writing code: 0.76 PM

However, I spent about 50% of my time on other things like science and technology intelligence, meetings and seminars, writing...

3- I m not sure I understand this point. If there's a collaboration, people should talk to each other. When I work on separate projects and there's opportunity to re-use code or other things, I do it. This is most efficient use of my time. Depending on the context, this may need to be disclosed to all parties but this has never been a problem before.

4- I don't deal with sensitive data so I can't comment on this aspect but having staff members on grants is possible. There can be all kinds of situations resulting from context, institutions and funders rules.

• 0 views • link
written by Jean-Karim Heriche Jean-Karim Heriche

Hi Jean – thank you very much for your response!

It might be good for me to mention that I may need change some things for myself (which is the topic of another blog post). Likewise, internal discussions are essential, and I have also found those to be very helpful.

I also want to make very clear that I am extremely grateful for all the support/guidance that I have received, and I expect that will continue to be crucial.

So, I will do my best to provide specific responses, each in a separate comment:

• 0 views • link
written by Charles Warden Charles Warden

1) It’s a long story, but I don’t think my full set of time-tracking data is a good fit for providing recommendations to others.

That said, I have listed what I thought might be helpful to others in this blog post. I also have some more details for response 2).

• 0 views • link
written by Charles Warden Charles Warden

2) It sounds like we have experienced a different need for follow-up requests, although something in the ballpark of 50% of time for communication / project discussion sounds about right for me. I think I am also trying to work on keeping education / literature review under 20% of my time, unless I substantially need to change something for myself.

I thought I heard that at least one other individual (from a shared resource facility?) encountered labs that realized that analysis would be more difficult than they expected (after they actually started a project). I am not sure how universal this is, but I think ~2 years is the length of time where I started to accumulate follow-up requests.

I think I should probably wait a little while to get additional feedback, but perhaps I can remove the current second question (which may be difficult for my peers to answer, if they have positions like myself) and replace that with something about follow-up requests (and trying to make expectations as clear as possible, at as early a stage as possible).

At least for RNA-Seq gene expression analysis, I have started to run some comparisons of public data to try and emphasize the need for methods testing for every project. That is described in this blog post which includes a links to the public data as well as the RNA-Seq template acknowledgements.

• 0 views • link
written by Charles Warden Charles Warden

It sounds like we have experienced a different need for follow-up requests

I work on a variety of project types, but only occasionally in bioinformatics. I normally consider a project open until publication unless it's very clear that my involvement is a one-off. Unexpected complications can happen and for long-running projects, I try to be pro-active in following up what's happening after handing over my part. This helps being prepared for follow up work. Several of my projects involve one or more iterations of data collection -> (exploratory) analysis -> new data collection ... which allows for the interleaving of projects I mentioned. Managing expectations is a must but goes both ways and when the project hits a snag this should be communicated as soon as possible to all involved. But from the initial question it seems more that the problem is more of a time management issue. At some point, if you feel overwhelmed by requests, you have to make it clear that requests are outpacing capacity. At this point, there are only three solutions: new projects wait until capacity becomes available, new projects get support elsewhere or the institution increases support capacity to match demand.
EDIT: Forgot to add: Don't compromise quality by rushing jobs to increase throughput. This is generally a bad idea as it often ends up costing more time later.

• 0 views • link
written by Jean-Karim Heriche Jean-Karim Heriche

Thank you again for your feedback!

I agree. I think I encountered a periods of being under-loaded after periods of being over-loaded, which I believe matches what you are describing. For example, you can be happy in the short-term, but then either be unhappy or discover problems that you didn't previously know about in the long-term. I am currently working with my supervisor to see what we can do in terms of avoiding being either over-loaded or under-loaded (although I may need to just focus on what I can change for myself, independent of any other general guidelines).

However, I am glad to see that you also have several projects where many rounds of analysis and discussion are necessary. I apologize for making it sound like our experiences may have been more different than they really were.

We have to be fair to everybody (and, if unexpected follow-up requests make grant and/or paper revision deadlines harder to keep, I think some sort of attempt at future needs may help). So, while a lab can prioritize, I think that could be a problem for a core. However, I very much agree that coordinating effort across the institute is important (although that may require the institute to agree something needs to be changed, at least in some situations).

• 0 views • link
written by Charles Warden Charles Warden

3) I don’t think this is the perfect solution, but perhaps I can give some sense of what I mean through acknowledgements:

The post-publication review section has a couple comments related to publication order. However, to be fair, I also want to emphasize that I have notes on post-publication review on papers where I am not an author as well as genomics companies (such as for human genomics and cat genomics companies). While it is a little different, I also had some thoughts on post-publication code support here (for a specific package).

• 0 views • link
written by Charles Warden Charles Warden

4) Great – thank you very much for your detailed response.

• 0 views • link
written by Charles Warden Charles Warden

Log in to answer this question.

More posts like this