The Silence (And The Grind)

I know, I’m late.

I haven’t posted in 36 days. I was planning on doing an update at Day 30, but honestly, I was too deep in a grind against OptKnock and my motivation was so down that I just couldn’t do it.

But I’m back! And I am ready to share the progress so far :D

The Learning Curve

For the past month, I spent a lot of time reading and self-teaching myself what Metabolic Engineering actually is. I learned from many resources and used AI to simplify the complex stuff.

It was long. It was tough. But I actually enjoyed it.

I read about 10+ papers (which is a lot to me! XD) to understand the tools and strategies researchers use. Some of the key papers that helped me:

• Designing Microbial Cell Factories for the Production of Chemicals
• Synergistic Production of Lycopene and Beta-Alanine Through Engineered Redox Balancing in Escherichia coli
• Production of Rainbow Colorants by Metabolically Engineered Escherichia coli (Really worth reading!)
• Metabolic Engineering of Escherichia coli for Natural Product Biosynthesis

The Work: Entering the Matrix

Once I learned the theory, I decided to get to work.

I looked up COBRApy and studied the documentation. I learned about Genome Scale Models (GEMs), Flux Balance Analysis (FBA), Flux Variability Analysis (FVA), and how to perform Knockouts for genes and reactions.

I dove deep into the E. coli core model to learn my way around reactions, metabolites, and genes. I learned how to find them, interpret what I see, how to add new ones and how to knock them out.

The Setup (In Silico)

Once I felt ready, I dove into the Lycopene Pathway.

I picked E. coli iML1515 as the GEM (since it is the most studied one) and jumped into the code. I introduced the Lycopene pathway, measured the theoretical maximum yield, and produced production envelopes.

Growth vs Lycopene Production

You can check the code for this phase here: Project Menhed

The Crash (Why I Disappeared)

Then I decided to try and achieve Growth Coupling myself… and it didn’t go as expected XD.

First, I tried single reaction knockouts. Then doubles. None worked.

I tried to use OptKnock and it wasn’t easy at all. I endured through it, but just as I got it running, it would almost always crash or hang for hours.

I figured out the free GLPK solver just isn’t fit for this level of math. I looked into Gurobi, but unfortunately, my university isn’t on the list for a student license.

So I decided to go the hard way: Brute Force. I used Python loops to manually knockout reactions in Single, Double, and Triple combinations to figure it out.

It was a dead end.

Either there is no solution, or I am doing something totally wrong.

The Pivot: Round 2

So right now, I am going back in.

But this time, I will follow the literature. I will try to replicate processes others have done to reach coupling. If that works, yay.

If not, I’ll adopt a Two-Stage Fermentation scenario for now and dive deep into the fluxes to see what I can do to make the cell produce the Lycopene.

By the way, I found out that in my announcement post, I promised to use AlphaFold. Turns out, I probably won’t need it since the enzymes used in Lycopene Production are pretty well known already XD.

Work continues.
See you in the next log. (Hopefully sooner this time! XD)