4 ms·

My interest is in applying bioinformatics methods to understand aging. His is bioinformatics in general. So most of the time there is actually no conflict betwe

by xaa 9y ago

My interest is in applying bioinformatics methods to understand aging. His is bioinformatics in general. So most of the time there is actually no conflict between "my" projects and "his" projects, my interests just being a subset of his. It doesn't hurt that my interests have introduced him to new people and funding mechanisms that he would not have otherwise had access to. I'm not sure how much detail you want. That was the short version. The slightly longer version is that there are huge datasets out there, and I want to understand aging by building tools and workflows for the many datasets which were not necessarily originally gathered to understand aging, but which still contain information about it. And also to work with and complement wet-lab aging researchers by giving them context or leads for experiments.

markkat 9y ago
That's fascinating. I started a company that cryopreserves adult stem cells. Would you mind sharing what kinds of datasets you are mining? Here or email (in my bio).
- xaa 9y ago
  Sure. The most mature part of this is analysis of expression data from GEO. We had to take the raw expression data and textual metadata, extract attributes from the semi-structured metadata into something statistically useful, and devise methods to do automated meta-analysis between experiments and platforms on the expression data. We are talking about >1M samples here by 20K-50K transcripts. Now we are basically extending the same idea to new data types and sources, like methylation and sequencing-based data. When we ask "what should we do next", we look for basically what archives are biggest, and if they conceivably have something useful to say about aging, we put it on the queue. The big-picture goal is to take this approach and apply it to all these different levels of organization (DNA, epigenetics, RNA, and ultimately hopefully protein) and put it all together in a cogent picture, using way more data than anyone has done before. We ask questions like "how does the aging process differ between tissues?". We also want to go "wide" by comparing species, as the whole "intervention X works in mice but not humans" is an important thing to address. Then on the smaller scale, the system is also of use and interest to individual investigators who want to know "what's happening with my gene X with age in tissue Y?". They like to include figures from us for example if they have some findings from mice and want to give evidence that the pattern may also hold in humans. Finally, we built the platform to be a pretty generic meta-analysis tool, so we have occasionally used it to answer questions completely unrelated to aging. Once you do all the background legwork, asking many different questions is largely a matter of changing the meta-analytic model.
jinto36 9y ago
When I read your first sentence my brain was briefly trying to figure out if I had sleep-posted to HN somehow because it sounded so similar to my interests (and there aren't a whole lot of people in the aging+bioinformatics area). I think I might have replied to some of your comments on other articles before but I seem to have lost access to the account I used to use. Are you working on just Human data right now? I have thought about trying to propose something very close to what you're talking about, but in the environment I'm in right now there's a pretty strong bias against non-hypothesis-driven work, and unfortunately the really big picture multi-omic approach is considered to be too vaguely defined. I wish I could just say "let's collaborate!" and then make it happen so I don't end up trying to duplicate what you're doing but if only the funding model actually supported that sort of cooperation (assuming you're in the US). If you don't mind, what institution/lab are you at now? I'm at University of Rochester, trying to make a co-mentorship arrangement work between a lab that works with C elegans as a model for aging (where the PI has no informatics experience) and a lab that works on computational modeling of biological networks (where the PI has no experience in aging/C elegans) because we don't have a real "bioinformatics" program. Also if you have any advice on bridging the gap between biologists and computationalists, that's something I have yet to master in my now more than half-decade of experience of trying...
- xaa 9y ago
  I'm at Oklahoma Medical Research Foundation. My focus has been on human, mainly because that's where the impact is, and because of minor technical issues making it easier to do human data. But we do other species as well, the system is just rougher around the edges for them. The way we get around the bias for (supposedly) hypothesis-driven work is to get a lot of our funding on collaborative and center grants. We can augment our local aging research work by providing them with context and pretty figures for their papers and grants, and they are happy to send a few $100K our way in exchange. We got a Shock Center award a year or two ago which has accelerated things a lot. A lot of it has been catalyzed by Arlan Richardson and Holly van Remmen, who came here from UT San Antonio. When it comes to "bridging the gap", yes, it is hard. The key I guess is to be flexible and understand what they want. A lot of them are not looking for grand, informatics-driven hypotheses, but they are ecstatic about extra figures or data they can put into their papers to spice them up. Sometimes I slum it a little and do routine statistical analyses on their experiments just to keep the connections live. In general, I think it is sensible to try to design systems in such a way they bridge multiple species from C elegans to rodents to humans. If you do one, it's not that hard to do them all, and they all have their place, as C elegans is good for quick testing of lifespan stuff but may or may not generalize, so it is considered helpful if you can use data to give tips about what may or may not translate. The best brief advice I can give is to figure out what the wet lab people want, and give it to them in a way that maximizes the overlap between what you are doing and what they are doing.

by xaa 9y ago

markkat 9y ago

That's fascinating. I started a company that cryopreserves adult stem cells. Would you mind sharing what kinds of datasets you are mining? Here or email (in my bio).

xaa 9y ago
Sure. The most mature part of this is analysis of expression data from GEO. We had to take the raw expression data and textual metadata, extract attributes from the semi-structured metadata into something statistically useful, and devise methods to do automated meta-analysis between experiments and platforms on the expression data. We are talking about >1M samples here by 20K-50K transcripts. Now we are basically extending the same idea to new data types and sources, like methylation and sequencing-based data. When we ask "what should we do next", we look for basically what archives are biggest, and if they conceivably have something useful to say about aging, we put it on the queue. The big-picture goal is to take this approach and apply it to all these different levels of organization (DNA, epigenetics, RNA, and ultimately hopefully protein) and put it all together in a cogent picture, using way more data than anyone has done before. We ask questions like "how does the aging process differ between tissues?". We also want to go "wide" by comparing species, as the whole "intervention X works in mice but not humans" is an important thing to address. Then on the smaller scale, the system is also of use and interest to individual investigators who want to know "what's happening with my gene X with age in tissue Y?". They like to include figures from us for example if they have some findings from mice and want to give evidence that the pattern may also hold in humans. Finally, we built the platform to be a pretty generic meta-analysis tool, so we have occasionally used it to answer questions completely unrelated to aging. Once you do all the background legwork, asking many different questions is largely a matter of changing the meta-analytic model.

jinto36 9y ago

When I read your first sentence my brain was briefly trying to figure out if I had sleep-posted to HN somehow because it sounded so similar to my interests (and there aren't a whole lot of people in the aging+bioinformatics area). I think I might have replied to some of your comments on other articles before but I seem to have lost access to the account I used to use. Are you working on just Human data right now? I have thought about trying to propose something very close to what you're talking about, but in the environment I'm in right now there's a pretty strong bias against non-hypothesis-driven work, and unfortunately the really big picture multi-omic approach is considered to be too vaguely defined. I wish I could just say "let's collaborate!" and then make it happen so I don't end up trying to duplicate what you're doing but if only the funding model actually supported that sort of cooperation (assuming you're in the US). If you don't mind, what institution/lab are you at now? I'm at University of Rochester, trying to make a co-mentorship arrangement work between a lab that works with C elegans as a model for aging (where the PI has no informatics experience) and a lab that works on computational modeling of biological networks (where the PI has no experience in aging/C elegans) because we don't have a real "bioinformatics" program. Also if you have any advice on bridging the gap between biologists and computationalists, that's something I have yet to master in my now more than half-decade of experience of trying...

xaa 9y ago
I'm at Oklahoma Medical Research Foundation. My focus has been on human, mainly because that's where the impact is, and because of minor technical issues making it easier to do human data. But we do other species as well, the system is just rougher around the edges for them. The way we get around the bias for (supposedly) hypothesis-driven work is to get a lot of our funding on collaborative and center grants. We can augment our local aging research work by providing them with context and pretty figures for their papers and grants, and they are happy to send a few $100K our way in exchange. We got a Shock Center award a year or two ago which has accelerated things a lot. A lot of it has been catalyzed by Arlan Richardson and Holly van Remmen, who came here from UT San Antonio. When it comes to "bridging the gap", yes, it is hard. The key I guess is to be flexible and understand what they want. A lot of them are not looking for grand, informatics-driven hypotheses, but they are ecstatic about extra figures or data they can put into their papers to spice them up. Sometimes I slum it a little and do routine statistical analyses on their experiments just to keep the connections live. In general, I think it is sensible to try to design systems in such a way they bridge multiple species from C elegans to rodents to humans. If you do one, it's not that hard to do them all, and they all have their place, as C elegans is good for quick testing of lifespan stuff but may or may not generalize, so it is considered helpful if you can use data to give tips about what may or may not translate. The best brief advice I can give is to figure out what the wet lab people want, and give it to them in a way that maximizes the overlap between what you are doing and what they are doing.