18 replies to this topic

[Illuminatusdarksoul]

hey guys,

I was wondering if you knew of any universities or labs that would want a second year bioinformatics student whose grades are averging between 65-75%. Until now i may be able to get a boring post doing some normal cell biology or something. What I really wanted is to use my skills to further the SENS cause. I emailed Aubrey de Grey, who was very kind and polite and suggested a topic, but I really need a lab somewhere and a supervisior to help me, and help me get the grant from somewhere.

If anyone knows a research lab that may take me this summer, or if indeed a research lab in Cambridge is working on SENS and is more staticly placed than someone who travels the world. I would be indebted as when i try to put my knowledge to use i am having difficulty doing so.

I have no idea whether Aubrey would have a summer placement or not, or whether giving a good research idea was his way of saying 'no sorry i cannot have u'. Should I give him another email in 3-4 days time

thanks 4 ur time.

stuart.
stu_bat@hotmail.com


thanks again guys!

Edited by illuminatus, 21 April 2006 - 04:15 PM.

[Mark Hamalainen]

If this is going to be your first lab placement then I wouldn't worry overly about getting a SENS placement. There is a very limited amount of funds for research directly related to SENS and not enough yet that it could be provided for basic lab training. I'd recommend taking one of the 'boring' posts and just doing as well as you can with the position so that you'll have a good reference when another opportunity comes around.

[th3hegem0n]

second year bioinformatics student whose grades are averging between 65-75%

So your GPA is like what, 1.0?

[Illuminatusdarksoul]

sorry i took so long to reply, my coursework was very time consuming and i have also being lacking of the internet.

[Illuminatusdarksoul]

I would like to say that i have taken a boring post, but on the side i asked aubrey about working on the 13mt proteins to see if i can make them more hydrophobic.

He asked me to write a program to do this, but i am probably only able to do the research manually- using other peoples programs (i am better at searching and biology than creating programs from scratch- a weak point in the bioinformatics field i know).

so my GPA is a first? well nearly, i am of first quality but a few areas usually drop me down to just under a first- i am rather mercurial and sporadic with my knowledge.

Can anyone help me out by telling me what the 13 mitochondrial genes are as this would get me off to a good start- i keep seeing people mention them, but so many people mention the 13 proteins that i cant find the damned things. not that i have looked to much- but i am trying to do this project to help you guys, as well as trying to do exams.

i apologise for not being on for a while- this is a most important site.

[Illuminatusdarksoul]

p.s the last time i emailed aubrey was 3 months ago about this project, do you think someone else has already done it??

[Illuminatusdarksoul]

the task is as follows-quote: "Allotopic expression
of the 13 mt-coded proteins from nuclear transgenes will, it's now very
likely, require changing the amino acid sequence to be less hydrophobic,
but of course one needs to do this without destroying enzyme function.
My guess is that this may not be as hard as it looks, because there is
no selective pressure either way on hydrophobicity of these genes while
they're in the mitochondrion -- they just need to be hydrophobic enough
to be stable in the membrane, but there's nothing to stop them becoming
"unnecessarily" hydrophobic. However, there is rather a lot of choice
out there in terms of which amino acids to change and to what -- rather
more choice than can practically be searched by experiment. Hence, the
bioinformatics task is to narrow it down. This is becoming easier every
week as more and more species have their mtDNA sequenced: all we need,
in order to identify candidates that may work, is to find positions at
which the human sequence has a particularly hydrophobic residue but a
lot of other species have less hydrophobic ones. This is not certain
to come up trumps of course, as there can easily be coevolution of the
gene and the nuclear genes for subunits that physically interact with
the subunit of interest, but it's a damned good start -- and in fact it
can probably be refined to prioritise candidates on the basis of the
phylogenetics of which residues are found in which species, so as to
identify ones that seem to change particularly easily and thus are not
likely to be sensitive to the coevolution problem." Aubrey De Grey


basically what i was going to do is look on NCBI for the 13 proteins, extract them in fasta format. then look for species that are near and far away from us, multiple aligning their protein sequences until i find more hydrophobic ones than ours, then do a write up on this.

though once i have found the amino acids in the sequence that can be easily changed to hydrophobic ones.- though to get around co-evolution i am a bit stumped as to how to do this, quantify it or which program to use.

p.s i am sorry if this seems to have wondered into the bioinformatics forum section. but i see the logic part of it and the start of the post as a biology post.

[kent23]

Can anyone help me out by telling me what the 13 mitochondrial genes are as this would get me off to a good start- i keep seeing people mention them, but so many people mention the 13 proteins that i cant find the damned things.

The 13


Okay, now please make a $13 donation to the mprize... Maybe you'll get to win it back someday...

[Illuminatusdarksoul]

Thankyou for that kent23, a nice link.

[John Schloendorn]

more hydrophobic

Ahh, since you mentioned this several times, please note that the job is to make them "less hydrophobic". I do like the plan though. It would be great if you could get things moving!

[Illuminatusdarksoul]

less hydrophobic. yes, i mean more hydrophilic- sorry, a typo. i think one thing, typed another.

[John Schloendorn]

though once i have found the amino acids in the sequence that can be easily changed to hydrophobic ones.- though to get around co-evolution i am a bit stumped as to how to do this, quantify it or which program to use.

Hmm, this is a question, isn't it? Perhaps if you added some grammar, people here might help...? What is co-evolution? If I'm guessing the question correctly, you could do something like:

- use a hydrophobicity table like this (but do get primary literature)
- make a "consensus" sequence of hydrophobicities parallel to your amino acid alignment (closely related species)
- make another "minimal hydrophobicity" sequence
- do consensus minus minimal, see where you get the largest differences
- use rational design to come up with suggestions, which aa of approximately the desired hydrophobicity would make the best functional substitute

[Illuminatusdarksoul]

Thankyou for your help John and others.

...right so if i made a concensus sequence and calculated the hydrophobicity of that sequence. (I can figure this out). But..

how would i create the minimal hydrophobicity sequence? I could maybe find the least hydrophobic sequence compared to the concensus sequence?

I was going to go about it by trying to find the exact a.a's from a large multiple alignment that could be changed to an less hydrophobic a.a.

In both cases I cannot think how to find out what is the least hydrophobic sequence is? I may just not be looking at this from the right direction.

Also I still do not see how this completely rules out co-evolution?

Edited by illuminatus, 23 April 2006 - 04:50 AM.

[John Schloendorn]

In both cases I cannot think how to find out what is the least hydrophobic sequence is?

Not sure where the problem is -- So I'll go through it from the beginning...

- The alignment is a sequence of letters, each representing an amino acid
- Each amino acid has a hydrophobicity value, as defined by the hydrophobicity table
- You can then actually align the hydrophobicity values (e.g. by using word's "replace" command on the aa sequence)
- You end up with an alignment block where columns represent the possible hydrophobicities each position can have
- Take the smallest value in each column (if you can't write a program to do this, do it by hand...), put them in a row, and there you have the minimal hydrophobicity sequence.
- If you get a considerable improvement, you can then translate it back to aa by looking at the original alignment, translate it into genetic code for mammalian expression and voilà, people can start synthesizing it, how fantastic would that be!

I think the art is to come up with the right stringency for the task. What if 10 species have ile to phe, and only one has arg? Is this for a reason? Should you play it safe and use phe or take a risk and use arg? Perhaps you can make a couple of different minimal hydrophobicity sequences based on the risk you are willing to take...?

Again, what is co-evolution and how is it a problem?

[Illuminatusdarksoul]

basically co-evolution is:

"coevolution is a change in the genetic composition of one species (or group) in response to a genetic change in another. More generally, the idea of some reciprocal evolutionary change in interacting species is a strict definition of coevolution."http://biomed.brown.edu/Courses/BIO48/27.Coevolution.HTML

this means that two species may differ in mitochondrial code for different reasons, and also if you were to put the genes in the genome aswell as the mt chromosome then you might get some unexpected interactions, or protein interactions could be changed. There can easily be coevolution of the
gene and the nuclear genes for subunits that physically interact with
the subunit of interest. - which basically means choose amino acids to swap that change easily, for any amino acid that we could use- as long as it is less hydrophilic (these a.a's will probably be less sensitive to any coevolution problem). So I need to look for a.a's that are more hydrophilic in other organisms, but are many different types of a.a's so as to try and prempt some sort of co-evolution problem.

p.s sorry if i was not clear about this- sometimes it goes over my head too.

[John Schloendorn]

choose amino acids to swap that change easily

I agree, this is the essence of the job, although I am not sure what this has to do with the above definition... but anyway, the problem you describe below seems to be why I suggested to come up with different risk levels, which you can actually quantify, to help the bench workers decide on what to bet their time and money.
And I do agree with you, there is no guarantee that this will work, let alone work easily. If you want to work for SENS better learn to live with it ;-) Do it, see what it gives you. I'm looking forward to first results! If you'd like to discuss more please don't feel discouraged by my occasional sarcasm.

[John Schloendorn]

Ahh, and so your question was how to quantify how easily a particular amino acid changes, is that it? For a start you might count how often it's same and how often is different and normalize by an accepted measure of phylogenetic distance such as ribosomal RNA sequence identity? But don't trust me, I'm just making this up... you will need to do some literature searches there -- someone out there should have had this problem before. Or perhaps someone here with a bioinformatics background can help?

[John Schloendorn]

How many species' sequences are there actually to do this with?

[Illuminatusdarksoul]

well most of the species such as drosophila and mouse are there, I am still looking for as many as I can. I am finding it time consuming to get all the latin names, but that is not a big problem, just a minor quirk in this sort of thing.

Yes thankyou for your help- I will look into how I could do quantify the phylogenies with regards to this.

Your sacasm is appriciated, though I was taking you more seriously, which is a mistake on my part- I am far too serious.

As far as phylogenetic analysis goes, to prempt co-evolution problems I think I may do a distant phylogeny minimum hydrophobicity sequence and one which uses sequences near to human. Better start compiling some trees.

Though as my exams are in 2 weeks I may put this on hold- cant have me failing at this point.