Eugene Myers on the Future Direction of Bioinformatics

Now that it has become rather inexpensive to sequence and assemble genomes, transcriptomes and metagenomes, the question is what to do with this advancement in technology. The conventional answer is to sequence as much as you can in various 1k, 10k and 100k genome/transcriptome/x-ome project fashion and ask questions later. This mentality led to abundance of GWAS (Genome Wide Association Studies) to determine ‘causal variants’ for all kinds of human traits. The approach failed despite inflated claims and media splash about dramatic discoveries, and the big claims were rightly criticized by knowledgeable scientists. For example, check the following two blog posts by Ken Weiss and Dan Graur -

The height of folly: are the causes of stature numerous but at least ‘finite’?

DisGWASted Again: Another #GWAS, Another Diddly Squat Accompanied by Loud Fanfare

Despite criticisms, the backers of GWAS and ‘more sequencing’ continue to claim - (i) the failure is due to small sample size, whereas their efforts will be successful with even more data, (ii) there is no alternative.

In a series of three posts (here, here and here), we showed that both claims were wrong. Borrowing from a paper by Sydney Brenner, we argued that GWAS and systems biology type of methods would not work due to flaws in fundamental scientific reasoning, whereas the alternative is to make cells, and not the genome, the center of attention. We also proposed a set of fundamental problems in biology and urged bioinformaticians to focus on solving those problems instead of building new tools for better genome assembly.

Speaking of developing genome assembly algorithms and other bioinformatics techniques, if one person had always been way ahead of time, that was Eugene Myers. In most core algorithm related to sequence alignment or assembly, you will find his significant contribution and he made those contributions in early 1990s ! He was the brain behind the company Celera, which was the first to sequence a human genome.

We sought opinion from Eugene Myers about our (and Brenner’s) suggestion about focusing on cells instead of the genome and he kindly replied -

I read the article and I pretty much concur. I work at an institute trying to understand development in terms of what cells can do, and I am dedicated to analyzing images and movies of what things encoded in the genome do in cells.

Let us add a bit more information to explain the above comment. Most bioinformaticians noticed that Myers took a break from sequence-related work since 2006-2007, right around the time when NGS took off. The conventional wisdom would say that he probably missed the trend, but those knowing his past record finds that he was again building the future way ahead of others. In 2009, his group worked with Stuart Kim to publish this interesting paper on C. elegans. This work brought forward Sydney Brenner’s work on C. elegans to the high-throughput age, and Myers’ group added the missing piece - computational identification of images of cells.

Analysis of Cell Fate from Single-Cell Gene Expression Profiles in C. elegans

The C. elegans cell lineage provides a unique opportunity to look at how cell lineage affects patterns of gene expression. We developed an automatic cell lineage analyzer that converts high-resolution images of worms into a data table showing fluorescence expression with single-cell resolution. We generated expression profiles of 93 genes in 363 specific cells from L1 stage larvae and found that cells with identical fates can be formed by different gene regulatory pathways. Molecular signatures identified repeating cell fate modules within the cell lineage and enabled the generation of a molecular differentiation map that reveals points in the cell lineage when developmental fates of daughter cells begin to diverge. These results demonstrate insights that become possible using computational approaches to analyze quantitative expression from many genes in parallel using a digital gene expression atlas.

More details are available here -

**Exploring Cells & Systems via Image Analysis and Customized Microscopy

**

We are best known for BLAST and the whole-genome shotgun protocol and assembler accomplishments in traditional sequence-based bioinformatics. But since 2002 the group has focused almost exclusively on analyzing and extracting information from images obtained by various forms of microscopy. We believe that such data will reveal more about the function of the entities encoded in the genome then any other approach and will eventually become a prevailing paradigm of investigation, like sequence-based discovery is today. The group has even begun to develop its own customized microscopes.

Since 2002 we have worked on the following representative problems among others:

Building a cell model of a C.elegans L1 larvae and software to automatically extract cell-by-cell expression levels in situ.

Building models of the fly nervous system using light microscopy. By stochastically expressing GFP in every individual neuron in this brain, we expect to deliver a detailed view of the highly stereotyped neuronal connectivity of the fly brain.

Tracking and measuring the vibrissae of a mouse while it is performing a cognitive function.

Tracking microtubules and centrosomes during the first few cell divisions of the embryogenesis of C. elegans.

Construction of an ultra-fast, block-face multi-photon microscope with onboard microtome for imaging a mouse brain with 1/2-micron pixels in less than a week.

In broad terms, the computational challenges are to build canonical 3D models of biological systems and map molecular observations onto the model. That is, there is the challenge of registering observations from different animals into a single representative framework, and the challenge of extracting meaningful information in the presence of low SNR and diffraction-limited resolution. The offsetting factor to low SNR and resolution is the presence of very strong prior knowledge about the morphology and dynamics of the entities under observation. Many of our new techniques thus involve what are called template- driven approaches.

We are in essence a technology group. Our aim is to develop microscopes and software that make observations of in situ and in vivo systems that enable our collaborative partners to advance molecular and cellular biology. We have worked closely with Stuart Kim (Stanford), Chris Doe (Oregon), Tony Hyman (MPI-CBG), Karel Svoboda, Gerry Rubin, Jim Truman, and Tzumin Lee (HHMI JFRC) as examples. The figure below illustrates imagery from three of our projects.

The overarching goal of our group is to build optical devices, collect molecular reagents, and develop analysis software to monitor in as much detail as possible the concentration and localization of proteins, transcripts, and other entities of interest within a developing cohort of cells for the purpose of working together with other groups in the Dresden area towards a biophysical understanding of development at the level of cell communication and force generation.