Pakistan’s insects occupy almost every ecological niche imaginable, whether it’s the grasshoppers of the country’s arid plains or the butterflies and moths of its forests, the aphids feeding on its crops, the beetles breaking down dead matter or the mosquitoes that can carry disease.

Traditionally, studying an insect meant placing it under a microscope, examining its physical characteristics and comparing it with existing descriptions. Today, that process is being augmented by DNA barcodes, genomic databases, machine learning, and artificial intelligence. The question has evolved from What insect is this? to What can its genes, images and data tell us that our eyes cannot?

Two species can look alike or a damaged or immature specimen may not have the necessary characteristics to help identify it. When physical traits reach their diagnostic limit, bioinformatics steps in to read the unalterable ledger coded inside the cells—the insect’s DNA.

Pakistan has already demonstrated what can be achieved at a national level through this method. A DNA-barcode survey of insect biodiversity examined material collected from 1,858 sites between 2010 and 2019. It eventually generated 50,592 barcode records, 49,363 of which were assigned to 6,590 Barcode Index Numbers (BINs).

Simply put, a BIN is a reference code used in the global Barcode of Life Data System (BOLD) database. When researchers analyse a short piece of an insect’s DNA, BOLD compares it with existing records and clusters similar specimens into the same group under a specific BIN code. This makes it possible to group insects that likely belong to the same species, even when their formal scientific name is unknown. This is how the BIN serves as a practical working proxy name for a species.

The study produced Pakistan’s first comprehensive DNA barcode profile of its insects, and the first survey of its kind in South Asia. But perhaps the most interesting finding was not what the researchers identified, but how much they couldn’t. Only 21 per cent of the BINs could be matched to species that already had a name in global databases.

At the time of the analysis, 59pc of the BINs had been recorded in BOLD, only from Pakistan. That does not necessarily mean these insects exist nowhere else. Much of Pakistan, as well as neighbouring countries, has simply not been sampled evenly.

In other words, we are still missing large pieces of the insect puzzle.

A barcode library can help make those gaps visible. It can show researchers where they need to look next, collect more specimens and, crucially, carry out the painstaking work of taxonomy—identifying, classifying and naming the species. And with researchers sharing data across borders, some of these unknowns may finally begin to get names.

Looking alike, behaving differently

Bioinformatics, which is essentially the use of science and math to store, analyse, and understand large sets of biological data, is crucial to deploy if the insect is an agricultural pest. So, for example, the whitefly Bemisia tabaci looks like one little white fly but scientists know that there is a complex of closely related species that are visually similar. These groups differ in how they spread plant viruses and respond to insecticides that seek to kill the bugs.

In one Pakistani whitefly study, specimens were collected from 255 locations in Punjab and Sindh and 173 records were added. The analysis resulted in the division of the material into 15 BINs, the identification of several putative species, and the identification of a previously unknown genetic group labelled “Pakistan”. Even though they may look alike, significant differences may be overlooked without DNA analysis.

A later study went beyond a short barcode and compared the whole genome of the Asia II 1 whitefly with a MEAM1 (Middle East-Asia Minor 1) reference genome. MEAM1 is a highly invasive biotype of the sweetpotato whitefly (Bemisia tabaci). It is globally recognised as one of the most destructive agricultural pests due to its ability to transmit crop viruses and rapidly evolve resistance to chemicals.

Researchers compared the genome of the Asia II 1 whitefly with the MEAM1 whitefly, using the latter as a genetic baseline. The comparison revealed about 2.33 million single-letter differences in the DNA, known as SNPs, along with another 202,479 insertions and deletions.

They identified variants in 14 genes previously associated with potential insecticide resistance. This is where bioinformatics comes in handy. Instead of trying to make sense of millions of genetic differences individually, computers can help researchers narrow the list down to a much smaller group of candidates worth looking into.

But a machine can only take the investigation so far. A genetic difference may look important on a screen without actually making a whitefly better at surviving an insecticide. That has to be tested in the laboratory and, ultimately, in the field. A separate study tested whiteflies collected from five districts of Punjab between 2017 and 2019. Their response differed according to the location and insecticide tested.

The lesson is simple. The computer can point researchers towards the suspects, but experiments establish whether they are actually guilty. Understanding insecticide resistance therefore requires all three pieces: bioinformatics to find the clues, laboratory experiments to test them, and regular field monitoring to see what is happening in real populations.

Bioinformatics is also being used to research the pink bollworm, another serious cotton pest. Scientists have used thousands of genetic markers to compare populations, including specimens from Punjab and Sindh. Another study used records from 17 districts of Punjab, laboratory observations and temperature-based models to estimate how infestations might change in the future. These predictions may guide researchers on where to concentrate their efforts, but they are not definite predictions. Field conditions such as weather, variety, planting time, and pest management can all have an impact.

What Pakistan can learn from global projects

The Anopheles gambiae 1000 Genomes Project (Ag1000G) is a collection of mosquito genome data in several countries across sub-Saharan Africa. This common information is used by researchers to monitor genetic variation, mosquito population movement and the spread of insecticide resistance. A similar network of universities, research institutes and provincial field teams is needed in Pakistan. This would enable the researchers to compare an insect collected in Faisalabad with the insects of Sindh, Khyber Pakhtunkhwa or Balochistan.

One example is the i5K initiative, which seeks to provide detailed genetic information for thousands of insects and related species. Pakistan can start with the key crop pests, disease vectors, pollinators and other beneficial insects. Every DNA record should be associated with a well-identified and curated specimen, and with basic information on its location and time of collection, and the host plant or animal species from which it was obtained. This would make it easier to review, compare and reuse information by other researchers.

Insects can, of course, also be misidentified, leading to the subsequent misidentification of their DNA record, which is why scientists don’t solely rely on bioinformatics. Contamination, incomplete databases, small samples and missing collection details can lead even the best astray. This is when entomologists exit the lab and descend into the field.

Bioinformatics cannot be a substitute for fieldwork, taxonomy, or lab experiments. Rather, it can unite them. It can reveal genetic differences between insects that look alike, highlight genes worth studying, and show how insect populations differ. Pakistan has already taken a good beginning. The next step is to enhance reference collections, exchange information, and recruit researchers in both entomology and computer-based analysis, and continue field sampling.

To look at an insect today is to see a dual identity: a living organism birthed by local crops and changing climates, and a precise record written in DNA. Pakistan’s scientific future depends on mastering both, learning to read the digital blueprint inside the cell without ever losing touch with the insect in the field.

Header image by Mohsin Alam

Trump made the proposal at his Republican Party’s first-ever midterm convention, but it was not immediately clear how the payments would work or whether they would be legal. With approximately 270 million US adults, the idea would cost around $1.35 trillion.

During a primetime address that lasted an hour, Trump boasted of completing “the greatest two years in the history of the presidency,” putting himself squarely at the centre of the election despite his deepening unpopularity.

But it was not clear whether his campaign-style speech would appeal to the critical swing voters who opinion polls show have drifted away from Trump’s party, amid growing dissatisfaction with the cost of living and the war on Iran.

“We hand America’s future to the radical left Democrats, and our nation will spend the next decade just getting back on its feet,” he said.

“Vote Republicans and we will leave the world in our dust for generations to come.”

The two-day Dallas gathering, where Trump will also deliver the closing remarks on Thursday after Vice President JD Vance’s keynote address, underscores the extent to which Republicans’ fortunes are tethered to Trump and serves as a showcase for their strategy ahead of the November 3 midterm elections.

The event was so centred on the president that Republican National Committee Chairman Joe Gruters informally dubbed it “Trumpapalooza,” a nod to the popular Lollapalooza music festival.

The focus on Trump amounts to a risky bet that he can mobilise the voters who carried him to the presidency two years ago, despite public polls showing his approval ratings have hit an all-time low. The president vowed to pay multiple visits to battleground states and districts in the eight weeks until Election Day.

“I’m asking you to pretend that I’m on the ballot,” Trump said, a line that may have pleased Democratic campaign strategists just as much as MAGA voters.

The convention presented a tough choice in particular for Republicans facing difficult elections who may need to distance themselves from the president to broaden their appeal but cannot afford to alienate Trump’s most committed supporters.

Many such Republicans skipped the convention, including Representative Tom Barrett of Michigan, Senator Dan Sullivan of Alaska and Senator Susan Collins of Maine, all facing strong Democratic challengers this fall.

Some, however, elected to travel to Dallas. At least 16 candidates in competitive House and Senate races were scheduled to speak, according to the RNC schedule.

Among those who spoke on Wednesday were Texas Attorney General Ken Paxton, a close Trump ally who is locked in a tough Senate race with Democratic lawmaker James Talarico, and Ohio Senator Jon Husted, who faces former Senator Sherrod Brown.

Though Trump insisted the 21,000-seat arena was “packed,” the upper sections remained largely empty during his speech capping Wednesday’s program.

Democrats portrayed as ‘socialists and communists’

While Republican officials had described the convention as an opportunity to tout Trump’s policies, most of the early speakers focused largely on attacking Democrats as extreme on topics such as transgender women in sports and border security—political red meat for the base.

Issues that voters have consistently cited as top concerns, the Iran war and the cost of living, received comparatively less attention.

Alabama state Representative Ernie Yarbrough, a convention attendee, said Trump needs to give voters a clearer sense of how long the war in Iran could last and what the administration’s aims are.

“People want to feel like we’re not getting stuck in an endless cycle,” he said.

Trump defended his decision to attack Tehran, saying he acted to prevent Iran from developing a nuclear weapon. He also promised that oil prices would drop “as soon as we win the war with Iran,” which he vowed would happen soon.

But the conflict shows little sign of ending more than six months after the US and Israel launched strikes on Iran, and analysts have said Iran’s nuclear program has not been eliminated. Iran denies seeking a nuclear weapon.

Before Trump took the stage, a parade of speakers called Democrats “socialists” and “communists,” echoing one of Trump’s favorite attack lines following a handful of victories by democratic socialists in Democratic primary elections.

Democrats are “dangerous, radical and weird,” said Senator Tim Scott of South Carolina, the chairman of the party’s Senate campaign arm.

Wednesday’s program also featured a surprise pretaped video from Democratic US Senator John Fetterman, who introduced his fellow senator from Pennsylvania, Republican Dave McCormick, as his “great friend” while declaring he would work with Trump to protect the state’s steel industry.

Fetterman’s extraordinary participation appeared likely to emphasise the ideological divides plaguing the Democratic Party ahead of the midterms. Fetterman, who has increasingly clashed with left-wing Democrats, has denied speculation that he may leave the party.

The convention’s stakes are high for both parties. With the elections less than two months away, Republicans face the prospect of losing control of at least one chamber of Congress, with public polling showing Democrats more motivated to vote and Trump’s approval ratings deeply underwater.

The event faced stiff competition for viewers with the start of the NFL season, with opening night coinciding with a Super Bowl rematch between the New England Patriots and Seattle Seahawks.

Still, after months of negative headlines over the Iran war and inflation, many Republicans see the gathering as their best chance to promote candidates, draw contrasts with Democrats and energise a Republican electorate expressing less enthusiasm about voting than Democrats, according to a Reuters/Ipsos poll.