• N&PD Moderators: Skorpio | thegreenhand

Is anyone interested in building a pharmacogenomics app?

CrimpJiggler

Bluelighter
Joined
Aug 28, 2011
Messages
241
I've built a system for interpreting raw data from 23andme and other genome sequencing services, I have the framework built, but to make a site like this successful, we need to focus in on a niche, a specialised area that the other sites won't get into. So the first niche that comes to mind for me is drugs. There are endless places we can go with this. Does anyone here with extensive knowledge in pharmacology, drugs, supplements, genes, medicine, biochemistry etc. want to partner up?
 
What kind of knowledge is requisite? Got dick in terms of programming, but good with drugs, pharmacology, mycology, botany, toxicology-(extensive, including common natural and synthetic toxins, exotic ones, modes of action, various vetinary differentials in susceptibility, fungal poisons, everything from lathyrogens and blindgrass to sea aenenome venom and potential medical uses of scorpion venoms in treating bran cancer [chlorotoxin, a fraction of the venom in the sting of Leiurus, the deathstalker scorpion, it selectively targets and binds to glioma cells, and is capable of giving a resolution many, many manyfold finer than that of MRI in detecting the buggers. My thought would be to prepare either a radiactive isotope-containing recombinant chlorotoxin or perhaps conjugated to a monoclonal antibody, labelled with a radioisotope pendant group, like how some ligands used to paint a tumour also include a radioactive metal, halogen, chalcogen or what have you.)

Also extensive knowledge of fungal toxins as well as those from plants, and venomous animals and both organic and inorganic chemical toxins toxalbumins (E.g abrin, ricin, curcin) plus various dinoflagellate and cyanobacterial products.

Especially where the fungi are concerned I've had to gain a good WORKING knowhow, in the practical sense since I've been hunting down the delicacies and treats to take home and munch on; in fact I, according to my grandfather was already possessed of quite the encyclopaedic knowledge of species and was able to explain for instance the mode of action of the likes of muscarine, cyclopeptide Amanita hepatoxins (amatoxins, virotoxins, phallotoxins and the protective, possibly, amanullin. This amatoxin is lethal in mice but the toxicity is species selective (its lethal in the murine model but not in humans, IMO there may be a chance of competitive inhibition of toxicity but I may be wrong, IIRC theres at least one of the amatoxins is protective against the rest. Other expertise includes neuropharmacology, biochem and, of course, drugs. Both their chemistry, their pharmacology, toxicity if any of significance, kinetics to a degree, subjective effects of some fairly exotic and little-tasted if at all used psychotropics.

For example, the brominated analog of chlormethiazole, hydroxymethiazole (the intermediate alcohol pre-haalogenation that is formed by cleaving vitamin B1, it seems to be active in and of itself although nowhere near as strong as chlormethiazole plus it tastes like absolute hammered shit, noxious, sulfurous, putrid stinker of a taste. Tastes as foul as ammonium polysulfides smell. Or for example, my recent experimentation with the truly strange alpha-chloromorphide.
 
If you were to sep up something like a chatroom for participants for this, such as one on tripsit (I think it is IRC), I'd consider applying to see what I could contribute. A larger pool of participants with multidisciplinary skills with not all of them being full-on project partner may be of use to you because of the wider scope possible and the 'peer review' 'second opinions'...

Am personally interested in pharmacology, drugs, supplements, some medicinie, biochemistry (studied chemistry, took bioinformatics but forgot a bit about it) and know plenty but am not a professional per se.
 
What do you plan to do with this site? Is it info only or applications in medicine and science? For consumers to use?

There are already sites that use known pharmacogenomic info, that can be customized for individual patients at clinics, labs, and pharmacies. So basically a professional can do a swab or other gathering technique, have it analyzed, and compared to known info to optimize drug therapy via pharmacogenomic info.
 
Thats what I've been thinking, if a group of say 10 of us come together and we all contribute our own unique areas of expertise, and our passions, we will rapidly create something this world has never seen. The tool I built is already the best genome interpretation service out there. I originally started with one partner, but he wanted to keep the team small, whereas I think we should have a big team of 10 people or so, everyone having equal equity and everyone being equally motivated to turn this site into something amazing, that this world has never seen. A system where we all prosper together, or we all suffer together. Limpet_Chicken seems to have extensive knowledge in toxicology, and toxicology will fit really well into this app because its important for people to know how genetically protected against toxin exposure they are. There is such a vast multitude of different detoxification genes, that the possibilities are endless.

My area of expertise is biochemistry, pharmacology and drugs. We need a nutrition expert, especially someone who knows both genes and nutrition. The first niche I want to target is drugs and medications. The framework for the site is built, we just need to specialise it, then we can launch V1 and start marketing it, that way we'll be one of the first sites doing this. I have endless ideas on where to go with this. Genomics is just the tip of an enormous iceberg, we can start adding blood/urine/stool tests, fitness tracker data, and questionnaires.

Where I think we should start is building things that bluelighters will find useful. For example, your genes determine how well various substances work for people. For example, codeine has to be converted by CYP2D7 into morphine before it becomes active. But it can also

If you were to sep up something like a chatroom for participants for this, such as one on tripsit (I think it is IRC), I'd consider applying to see what I could contribute. A larger pool of participants with multidisciplinary skills with not all of them being full-on project partner may be of use to you because of the wider scope possible and the 'peer review' 'second opinions'...

Am personally interested in pharmacology, drugs, supplements, some medicinie, biochemistry (studied chemistry, took bioinformatics but forgot a bit about it) and know plenty but am not a professional per se.
 
For example, codeine has to be converted by CYP2D7 into morphine before it becomes active.

it's CYP2D6 that demethylates the opium alakaloids, and no, codeine doesn't need to be metabolized via that route to be effective as an analgesic - evidence shows that the codeine-6-glucuronide produced via the UGT2B7 enzyme is a larger contributor to the analgesia. It's also probably the same case with dihydrocodeine, judging by homology and the relatively low relative conversion of DHC to DHM in vivo. It's also been proven that oxycodone and hydrocodone are active on their own and any metabolism via demethylation plays but a minor role in the analgesic effects.

how genetically protected against toxin exposure they are

Whatever happened to "dose makes the poison"? What a nebulous statement... "protected against toxin exposure".

BTW, does this project hinge on expecting people to have intimate knowledge of how individual genes effect the human phenotype? From the way you're describing it it certainly seems that way. I can see that being a task fraught with peril - you may have a large database of SNP's but how on earth do you plan to accurately correlate them to drug effects while excluding the effect of set and setting? I expect otherwise you may just have a lot of inferences derived from Wikipedia articles on genes.

Exactly how much data is provided anyway? I thought 23andme did just SNP testing? Entropy would seem to be working against you then, in that you can only infer so much if you only know the status of a few bases per gene... and presumably some genes lack information at all. Does it also include the mitochondrial DNA as well, I wonder?

Good luck on your endeavor in any case. Let me know if you're willing to provide more concrete payment - I will repeat here that I don't like undertaking big projects for free... on the lawless frontier of the internet, working for "equity" means dick - and the guy who has control of the servers with the project on them can cut and run whenever he likes, leaving all the contributors in the dust. I've seen that happen too many times - that or the whole thing loses momentum.

Have you considered open source development? You may be a little more effective in recruitment that way. Also, there seems to be stiff competition in this field.
 
https://www.pharmgkb.org/pathway/PA146123006
Morphine is way more active than the glucoronide. Way more glucoronide is produced than morphine.

Whatever happened to "dose makes the poison"? What a nebulous statement... "protected against toxin exposure".
Whether a particular dose is a poison or medicine/inactive, is influenced by ones genes. Thats the idea, using knowledge of those genes to get a better idea of how well a person can tolerate a particular compound. Benzene for example, it'll be a lot more toxic if your NQO1 gene isn't functioning optimally.

BTW, does this project hinge on expecting people to have intimate knowledge of how individual genes effect the human phenotype? From the way you're describing it it certainly seems that way. I can see that being a task fraught with peril - you may have a large database of SNP's but how on earth do you plan to accurately correlate them to drug effects while excluding the effect of set and setting? I expect otherwise you may just have a lot of inferences derived from Wikipedia articles on genes.
There are vast resources available. There are thousands of GWAS and studies correlating genotypes with ailments and drugs. Theres no way to factor in set and setting, but thats not really relevant, gene influences will be there regardless of set and setting, unless you're talking about epigenetics. I believe set is an epigenetic factor, that things like beliefs can influence expression of genes, so if thats what you meant, I'm with you there. Until methylation and aceylation pattern sequencing becomes readily available, we're gonna have to leave epigenetics out of this project. But my god I can't wait for the day that we can build apps to interpret methylation patterns, imagine the possibilities.

Theres near endless possibilities for what to do with a site like this, and personal genomics is going to explode over the next 5 years or so.
 
Exactly how much data is provided anyway? I thought 23andme did just SNP testing? Entropy would seem to be working against you then, in that you can only infer so much if you only know the status of a few bases per gene... and presumably some genes lack information at all. Does it also include the mitochondrial DNA as well, I wonder?
It depends on what chip was used to sequence the DNA. 23andme V3 has 600,000 SNPs. The chip before that had a million. There are other sequencing services like ancestry.com which have their own chips. And yes, 23andme includes mitochondrial DNA.

Also, there is no stiff competition. All the sites on that list do different things to what we do. Equity doesn't mean dick, unless we setup an LLC. I spent a year on this project so far, so I won't be cutting and running, I plan on working on this for life.
 
Top