You Too Chitragupta!

I was standing in front of the office of Chitragupta (CG, henceforth), the moral bookkeeper of good and bad deeds according to Hindu scriptures. CG was somber like customs officers at the US entry ports and received applicants of the afterlife with no customary smile. In a soundproof glass chamber, he dealt with each of the recently deceased human souls with a sense of detachment. After a thorough examination, he was passing his judgment by stamping a seal on their heavenly passports, barely visible from the lounge where I was waiting for my turn. However, the expressions on the faces of the applicants were revealing loud and clear the color of the stamp that was being punched on the passport. Every bright green ‘accepted’ stamp would yield cheerful shouts, with the applicant disappearing on their path to heaven. It was evident when the red ‘rejected’ stamp was printed on the passport as the applicant said one final cuss word on their way to hell.

When my turn came, I tried to smile, not wanting to upset the mighty officer in God’s office. CG looked at me even as the biometric machine confirmed it was really me who had died and not my lookalike. He got to the real business after the customary exchanges establishing my credentials.

CG: Your primary role on the Earth?

GB: Academician.

CG: You will have a choice of selecting your best research contribution.

He brought up a deck of my printed research papers with a click of his thumb and said, ‘Pick just one!’ while pointing to a large trash can labeled ‘Never Mind.’

I reminisced about my life as an academician. Looking back, I tried to identify the most enriching research explorations I have indulged in the last two decades as an active researcher by weeding out those I consider not-so-worthy. Of course, only in retrospect. Then, each of them seemed equally valuable because of its intriguing nature or due to its potential value. No doubt, I have been occasionally driven by the peer-perceived importance of the research or the impact (factor) of the venue where it might land. Thankfully, such instances have been far few because of self-reflection-induced nausea.

After some initial hesitation, I could throw most of my published research into the ‘Never Mind’ bin except for ‘three.’

CG: Three is two too many? You gotta throw two of these.??

I went into a flashback mode. Chronologically, the first of these three treasured studies was about modeling the network of cities interconnected via airlines. Analyzing India’s aviation infrastructure as a complex network is among my most cherished studies that I completed as a graduate student. From the conception, hunting for data at railway stations and reception desks, coining the term Airport Network of India, its analysis, manuscript writing, seeking feedback, to publishing the preprint on arXiv, it was one heck of an adventure. As one of the earliest characterization studies of aviation infrastructures, it remains among the well-cited articles of mine... ‘Analysis of the airport network of India as a complex weighted network,’ G Bagler, Physica A: Statistical Mechanics and its Applications, (2008).

Nonetheless, I was compelled to ‘trash’ this research of mine. And so I did.

CG: That still leaves two of your research articles on the deck!! Pick just one.

The mood in the chamber changed again, and I once again went into retro mode.

The next one was the core of my PhD thesis—the discovery of ‘assortative mixing’ in graph-theoretical models of protein structures. This study originated in my efforts to break down some of the earliest articles that used networks as a paradigm to model the native state protein structures. Besides reproducing some of the early research in this direction, we investigated topological properties of ‘Residue Interactions Graphs (RIGs),’ which captured coarse-grained details of protein structures. Wading through the analysis using an elaborate FORTRAN90/95 code led me to the exceptional assortative mixing in the RIGs, a generic property ubiquitously present across all globular proteins!

Enamored by this universal phenomenon, I had a gut feeling that this property of protein structures must have a basis in their functions. I remember presenting my early results in STATPHYS-KV, a biannual statistical physics meeting held at Satyendra Nath Bose National Centre for Basic Sciences, Kolkata, in 2004. One elderly gentleman approached my poster and asked many probing questions about the results. Soon, he returned with another lady and pointed to one particular panel with a dint-I-say-so look. That part of the poster demonstrated the assortative mixing in eighty protein structures across structural classes. I later learned the names of these senior physicists from PK, a postdoctoral fellow: Deepak Dhar and Indrani Bose!

Knowing the potential importance of the result, for months, I unsuccessfully attempted to relate the ‘folding nucleus’ of proteins to the amino acids that form the ‘assortative core.’ Expectedly, the unusual assortativity in protein structures I had observed was brushed aside by my biologist friends as a Physicist’s fascination for patterns with no meaningful ‘biological implications.’ Even as I struggled to find meaning in this topological quirk of protein networks, I went to Beijing to participate in Santa Fe Institute’s 2005 Complex Systems Summer School, where the magic happened.? ??

I was gung ho about my observation of generic assortativity in globular proteins. I took every opportunity to share my excitement with anyone willing to listen, from Cosma Shalizi, Geoffrey West, and John Holland to Michelle Girvan. After Michelle’s talk, I approached the dias and told her about my observation. I tried quickly explaining my discovery in the noisy lecture hall with words like proteins, networks, assortativity, and the folding nucleus. And she was like, ‘Oh! So you mean assortativity is correlated with the folding rate.’ I remember not saying a word after that as I had immediately steered towards Michelle’s ‘folding rate’ hypothesis.

Since we had limited access to the internet in the remote resort location, Xiangshan Park in The Fragrant Hills, where we were putting up in the first week of School, I had to wait until we moved to the Chinese Academy of Sciences campus in Beijing. Imagine my ecstasy when I got hold of previously published research that reported a negative correlation of the experimentally reported folding rate with simple metrics that capture long-range interactions. None of these invoked the network representation of proteins or metrics as nuanced as ours. Notably, the assortativity coefficient positively correlated with the rate of folding compared to the negative correlations reported thus far. With such a headstart, I could complete the rest of the investigations to publish this result in the Bioinformatics journal of Oxford University Press: ‘Assortative mixing in protein contact networks and protein folding kinetics,’ G Bagler and S Sinha, Bioinformatics (2007) .

Protein structures have been a matter of investigation for decades. The earliest study with graph theoretical representations of proteins goes back to 1976. Hence, discovering an exceptional, generic phenomenon in globular proteins by invoking the graph theoretical paradigm for modeling biological macromolecular structures and pining it to its biophysical property of folding kinetics is a cherished moment in my academic journey.

Alas, I have to pick just one! Therefore, this research, too, goes down the drain. I reluctantly dumped this too in the ‘Never Mind’ bin, leaving me with one last research article.

I pointed to my choice, suggesting that this was my final choice if I selected just one.

CG: Are you sure?

GB: Of course!

This final research conducted in 2014-2015 was my last academic milestone, chronologically speaking, that I am proud of—A rigorous investigation of food pairing patterns in Indian cuisine. ‘Spices form the basis of food pairing in Indian cuisine,’ A Jain, NK Rakhi, and G Bagler, arXiv preprint, arXiv:1502.03815 (2015). ?

Sitting in my lab at IIT Jodhpur’s makeshift campus in Ratanada, I remember designing experiments to test the food pairing hypothesis for Indian recipes. After carefully reviewing many data sources, we zeroed in on the online repository of recipes by Tarla Dalal, the legendary cookbook writer in modern India. Our investigation of over 2,500 Indian recipes from regional cuisines across the subcontinent led to the exciting discovery of a unique ingredient pairing pattern and the central role of spices. MIT Technology Review highlighted this research in its Emerging Technology section. That opened a floodgate of coverage by the global media, including the Washington Post, National Public Radio (NPR), The Hindu, and Times of India. On the 4th of March 2015, I remember my laptop was teeming with messages and beeping with fluttering sounds as this news started trending on social media.?

Eight years down the line, much has happened.

What started with a serendipitous exploration, the study of food through the lens of data science and computing, has a globally accepted name—Computational? Gastronomy. What began as a humble, curiosity-driven question in a classroom in the makeshift IIT Jodhpur campus in Ratanada has become a new way of thinking about food. From an acute scarcity of food data in 2014, today, we stand in the middle of structured data repositories on traditional recipes from across the globe, nutrition, flavor, health associations, and carbon footprints. Our lab, the complex systems laboratory , is the ground zero of computational gastronomy. IIIT-Delhi has launched a first-ever course on Computational Gastronomy. Over 200 research interns have contributed to the edifice, which has also led to the creation of a start-up.

I was pretty sure this was one article that will sail me through to the gate of heaven!

CG: This is your last chance to take the call. Are you confident?

GB: Yes.

CG: Should I lock it?

GB: [visibly irked by the Bachchanishque KBC style] Go ahead and just do that!

With tongue firmly in cheek, CG punched my passport with a bright red stamp, showing me the way to hell with his left hand.

Stunned as I was, I still read through the fine print of the REJECTED application.

It read as follows: ‘Preprint. No peer review. Less than 100 citations.’

I was livid, angry, and woke out of the nightmare murmuring—

‘You too, Chitragupt!

Ganesh Bagler